Wikipedia talk:Criteria for speedy deletion/Simplify policy RfC

From Wikipedia, the free encyclopedia
Jump to: navigation, search
The following discussion is closed. Please do not modify it. Subsequent comments should be made in a new section. A summary of the conclusions reached follows.
The following views are strongly supported by consensus:
  1. CSD must be kept clear and specific, bright-line and mechanical. [1.8]
  2. Out-of-CSD speedy-deletions should instead be passed on to XfD (or PROD) for discussion, and non-compliance with this should be very rare.[1.3]
  3. The CSD do not oppose IAR, and IAR-like exemptions should not be mentioned in the CSD.[1.3,1.6]
  4. Strict CSD criteria are not bureaucracy, but instead represent the consensus against undiscussed deletions.[1.3]
  5. Though admins usually have good judgment,[1.4] higher standards apply to CSD due to less oversight,[1.8] and due to the harmful effects of deletion. Loosening the standards will make the effects worse.[1.3,1.11]
  6. Deletion, especially by one party, often has the serious harmful effects of upsetting and discouraging new contributors who are acting in good faith.[1.3]
  7. Our best-practices may be prescribed to administrators, though the descriptive/prescriptive distinction is not actually that significant an issue.[1.6,1.10,1.11]
  8. (There is some support for refactoring CSD [1.2,1.9])

Background: The Wikipedia:Criteria for speedy deletion page contains the speedy deletion policy. There is a discussion as to whether this policy should be made less prescriptive and more descriptive. 16:00, 20 May 2009 (UTC)

Proposed changes
  • Alter the language of the policy, especially in the introduction, so that it does not imply that the policy is a firm rule;
  • Generalize the criteria where necessary so that administrators making deletion decisions are not bound by technicalities, but instead are bound by the spirit of the policy;
  • If necessary, further simplify the policy by reducing the number of criteria or merging them, so that the policy is easier to follow and the spirit of the policy is not obfuscated by jargon.
Possible objections to the proposals
  • The policy has been this way for a long time and therefore it already reflects community consensus. So there is no need to change it.
  • Policy that is not prescriptive is merely tactical information, and as such it would not be policy at all. Therefore the policy is prescriptive or it is not policy.
  • Administrators do not need leeway to interpret the criteria, since any ambiguous content can be sent to WP:XFD.
  • The criteria already represent consensus for what counts as a valid speedy deletion, and so an administrator who deletes outside the policy necessarily contravenes consensus.
  • When administrators are given too much leeway, they are bound to disagree. A clear and unambiguous explication of this policy is needed to resolve these disputes.


Please add and sign your statements below.

Statement by Ryan Delaney (talk · contribs)[edit]

The fundamental policy of Wikipedia is this: "If it improves the encyclopedia, do it. If it doesn't, don't. If you aren't sure, look it up or ask." Wikipedia policy pages are descriptive, not prescriptive; and they are subject to the occasional exception. The project-wide consensus supporting these principles is nearly universal, but the opponents of this proposal argue that an exception should be made for this policy, and that it should be considered prescriptive. As such, I view this policy as it is currently written, and the consensus supporting it, as a referendum on the existing core policies of Wikipedia. One of the two will have to give way, and I think it should be the former. The five pillars explicitly repudiate interpretations of policy as literal and prescriptive for very good reasons, and those reasons obtain in this case as well.

A literal interpretation of this policy simultaneously gives administrators too much and too little power:

  1. Too much power, because an administrator can speedy delete any page that could be interpreted to fit one of these forty or so criteria and then hide under the policy. As such the CSD policy crowds out critical thinking and leads people to believe that they can justify their (wrongful) actions by citing policy rather than explaining why what they did helped the project.
  2. Too little power, because articles that obviously and unambiguously fit the spirit, but not the letter, of the CSD policy cannot be speedy deleted. This literal interpretation of policy leads to severe instruction creep since people think they have to pack the policy with everything that they might ever want to speedy delete: non-notable corporations, bands, pets, street signs, screwdrivers, etc. This has led to an endless bloating of the policy, with more critera, prescriptions, caveats, and assorted details being added and removed, which every administrator is expected to memorize and follow to the letter.

By reorienting our attention away from blind adherence to rules and back onto individual case-by-case editorial judgement, we are making administrators responsible for their actions, instead of giving them a blanket to hide under whenever they speedy something that might fit one of the criteria. Everyone should be ready to explain their administrative actions with an explanation that is based on why it was better for the encyclopedia, not why it was consistent with a policy page. The Wikipedia:Criteria for speedy deletion policy should reflect this, both in the way we interpret it, and in the way we write it. To this end, the proposed changes should be implemented.

--causa sui talk 19:47, 19 May 2009 (UTC)

Users who endorse this statement[edit]
  1. Excirial (Contact me,Contribs) 14:24, 20 May 2009 (UTC)
  2. Mr.Z-man 15:18, 20 May 2009 (UTC)
  3. --Elen of the Roads (talk) 18:31, 20 May 2009 (UTC)
  4. Shadowjams (talk) 07:51, 21 May 2009 (UTC)
  5. hmwithτ 11:31, 24 May 2009 (UTC)
  • The present policy describes what good admins do , and requires that all do things that way. The question is not prescriptive/descriptive, but restrictive/permissive. DGG (talk) 14:53, 20 May 2009 (UTC)
    Making the distinction "restrictive/permissive" is to have assumed that the policy is prescriptive, since neither of those options makes any sense outside of that dubious assumption. The policy does require that all admins do things in a particular way, and that is exactly what is being challenged in this RfC. --causa sui talk 15:05, 20 May 2009 (UTC)
    I agree with DGG here. Bearian (talk) 18:41, 21 May 2009 (UTC)
  • It's not true that if a page doesn't striclty fit into one of the CSDs, it isn't speedy deleted. For example, some users (including me) have deleted self redirects before CSD R1 explicitly covered them. עוד מישהו Od Mishehu 07:47, 21 May 2009 (UTC)
    Good on you. --causa sui talk 08:05, 21 May 2009 (UTC)
    It is, however, true that some people, including many of the people most active on the CSD talk page, would very much like it to be true and some treat it as true. Mr.Z-man 22:00, 21 May 2009 (UTC)
    It is, however, true taht some people, particulalry those who never participate on CSD/Deletion discussions, are the ones whom I have the least amount of respect for. (I say this namely because the two admin's whom I'm the most critical of and have thought about starting RfC's on, I've NEVER seen on any talk page.)---I'm Spartacus! NO! I'm Spartacus! 00:33, 24 May 2009 (UTC)
  • erm... OK, I admit that I'm not active in any XfD stuff. Other than some !votes, and my opinions during the move from 5 to 7 days, I just don't do much in this line. Not that I don't value it, I know it's important and needed - it's just not something I've been involved in. Hopefully I'm not sticking out like a sore thumb here - and indeed, I'm not even sure where to post. My thoughts are this: When I run across an article that needs deleted (in my opinion), say when I'm doing recent changes or NPP - by the time I figure out which CSD tag should be on it - it's already been deleted. If the criteria were a bit more simple, I'd be a bit more inclined, or rather less hesitant to tag an article. Just my thoughts - so given my efforts in this area - judge the proper "weight" with it as well. That is all. — Ched :  ?  03:08, 3 June 2009 (UTC)

Statement by WereSpielChequers (talk · contribs)[edit]

CSD is a complex policy which is probably interpreted differently by different admins and new page patrollers, and would benefit from a fundamental review. However the group of editors who we should be most concerned about when discussing CSD are the authors of these articles; Often these editors are newbies whose first encounter with Wikipedia is to find their article summarily deleted within minutes of them saving the first sentence.

Currently we allow any newly registered account to create new articles, without warning them that their first edit has to create an article that can survive the speedy deletion process. This could be fixed by setting a little bit of code that automatically created new articles by newbies as sandbox articles in their user space - differentiating between saving drafts of an article and submitting it to mainspace.

One advantages of different codes is that different categories of speedy deletable articles generate different templates for posting on the author's talk page. We need to give very different messages to the writer of attack pages than we do to the person creating an article on their village cricket team. Different codes enables us to generate appropriate messages both for the article and for the authors talk page.

Currently communication with the author is considered to be optional. Whilst it would be overkill to give an editor eleven templates when you tag all eleven members of their village cricket team for deletion, if a new editor with a redlinked talkpage has had articles speedy deleted something has probably gone wrong (perhaps a bot could resolve that one).

For some deletion categories {{Move to my sandbox}} should be offered as an alternative to {{hangon}}.

The prod and speedy tags should be merged and rationalised. The new page patroller shouldn't need to know whether the tag they are going to apply will result in deletion within minutes, hours or days; they just need to know whether the article they are tagging is an attack, spam, vandalism, hopelessly non-notable etc etc.

NoContext should not be a criteria for speedy deletion as most articles tagged as no context are salvageable with a little bit of work - it is a useful tag, but should not be treated as a speedy deletion criteria but should be a prod rather than a speedy.

Totally non-notable articles such as biographies of pet rabbits, "the most awesome guy in the world" and 12 year old political activists should get a slightly different message and less time to add detail than subjects who are notable enough for their parish newsletter but not for here.

ϢereSpielChequers 00:13, 20 May 2009 (UTC)

Users who endorse this statement[edit]
  1. Agreed par my comment below. Excirial (Contact me,Contribs) 13:15, 20 May 2009 (UTC)
  2. Unionhawk Talk E-mail 18:23, 20 May 2009 (UTC)
  3. Although it's not what's under discussion, it is a valid statement--Elen of the Roads (talk) 18:30, 20 May 2009 (UTC)
  4. I could support the move to create a time-delayed CSD category. My biggest concern is the speed at which some people arbitrarily delete articles within minutes of being created.---I'm Spartacus! NO! I'm Spartacus! 00:51, 24 May 2009 (UTC)
  5. Ched :  ?  17:03, 31 May 2009 (UTC)
  6. Largely support. Lots of good ideas here, but I'd tweak some things; for instance, "the most awesome guy in the world" shouldn't get less time for improvement, it should get no time, since stroking vandals and self-promoters has been shown to be counterproductive. - Dank (push to talk) 14:01, 11 June 2009 (UTC)
  • I find this a very interesting point of view Spiel. One of the issues with new users is that they do not know their way around Wikipedia yet. Hence, some of the first edits are new pages. Most users likely don't even know how to challenge deletion or request explanation at a talk page. Even if they do there is most times a need to quote an entire squadron of rules (WP:NPOV, WP:ADVERT, WP:N, WP:V, WP:RS, WP:CORP and on and on) to explain them what is wrong with the page. Even when trying to explain these manually it leads to an information overload - and subsequently to a confused user. I know there is a page somewhere that reviews articles created by IP users for inclusion. What about sending articles made by very green editors here for review, rather then bombarding them into the article space immediatly were many face CSD? Excirial (Contact me,Contribs) 13:06, 20 May 2009 (UTC)
    I agree and that is part of what I'm trying to suggest, however there needs to be a process whereby they spot that we don't currently have an article on x before they start it and then find their article is automatically in a "new articles by new writers improvement zone" - otherwise we will have thousands of alternative articles on contentious people. A sandbox in their user space would be one way to do this, alternatively as part of the flagged revisions process this could be an article not yet cleared for inclusion. In my view we should be helping these newbies not scaring them off, and if that means a not quite notable article hangs around for a few days outside of article space so be it. ϢereSpielChequers 13:36, 20 May 2009 (UTC)
  • It's an interesting idea, but I have to say that this is not really the topic of the RfC. Though I'm open to the idea of simplifying deletion by merging PROD with CSD, I want to make clear that what's at issue here is reform of the CSD page content, not deletion process in general. If this proposal gets a lot of support, we might start a separate discussion for it elsewhere? --causa sui talk 13:37, 20 May 2009 (UTC)
    • Thanks Ryan, I apologise for going somewhat off topic in my comments, but I started off responding to your point "If necessary, further simplify the policy by reducing the number of criteria or merging them, so that the policy is easier to follow and the spirit of the policy is not obfuscated by jargon." My concern being that merging and simplifying should not be done at the expanse of communicating less clearly with authors, especially newbie authors. However a simplification such as widening A7 from "An article about a real person" to "An article about a real being". Would retrospectively legitimise the one time I've deleted a pet under that criteria. ϢereSpielChequers 10:19, 22 May 2009 (UTC)
  • Many of these suggestions are good (some I have long advocated), and each proposed modification should be considered seriously, though I would certainly oppose merging prod & speedy. That;'s not the current proposal, though. DGG (talk) 14:58, 20 May 2009 (UTC)
    Well, i think that a merger might actually be a good thing for both the deletionist and inclusionist camp. the PROD template is mainly used to remove pages after a set amount of time if there is no reason to remove a page with all due haste, with the annoyance that anyone may remove it. A CSD is more restrictive and may not be removed by the person who placed it but articles are often removed in a matter of minutes or hours.
If we merged the two we would get something akin of a delayed CSD. It means the article would get 7 days to be improved (Unless it falls under G10, personal attack) before it will be removed trough CSD, but it will deny the right to remove the template at will. Most CSD pages don’t have a specific need to be removed at once, and most prod removals are by page creators who don’t change the article. We could kill not two, but three birds with one stone: The annoyance of using PROD templates, and the “Unfriendly” to fast removals, and allowing for the article rescue squadron or other inclusion focussed groups to check and improve pages marked for removal. It is just an idea, but would this sound reasonable to you? Would there be any drawbacks that I miss? Excirial (Contact me,Contribs) 18:45, 20 May 2009 (UTC)
I think what you want is a Speedy tag that can only be used on items that are a significant risk to the project (copyvio etc, see my statement below), and a slightly more 'sticky' PROD that can't be removed by the content author. If the article has not been improved in any way at the end of the PROD period, it can go. If it has, then the PRODding editor should take it to AfD if they still think it's bad.Elen of the Roads (talk) 18:54, 20 May 2009 (UTC)
Along those lines indeed. The entire "G" deletion category for deletion lists only contains issues that need immediate handling (Attack pages, copyvio's) or issues that are non controversial (Empty articles, blatant spam). Those can be speedy deleted without a second thought. The articles in the A category judge content - if we reside to that that it likely means there is nothing so wrong with the article it will have to be removed within minutes. The first category could fall under speedy, while the rest could receive a prod that a user could only contest instead of remove.
Still, i see two problems with this: The first one is that a delayed speedy is still 7 days. If it is removed in between it should be re-added automatically or we will still have to check each article that. We could try an abuse filter but what would happen if an alternative account or IP removes the template? The only way to counter that is trough disallowing everyone except admins to remove the template which is probably a bad idea. The second problem is where we draw the line. Should we sticky prod an article that says "Born: 7/2/1992 Age:17 With a creative mind, and a budding career at the Burren Center alison is bound for Stardom" and wait 7 days or should we just remove these trough CSD? And if we remove trough CSD, wouldn't that completely destroy the entire sticky prod idea? Excirial (Contact me,Contribs) 19:51, 20 May 2009 (UTC)
I doubt there's a perfect solution to this. I think that any admin has to have discretion to "delete, this is going nowhere" - given that any deletion can always be challenged.Elen of the Roads (talk) 21:02, 20 May 2009 (UTC)
Blatant Spam is one of those problematic categories... most of the articles I see labelled blatant spam are spamish, but not necessarily so. The line between a blatant spam article and one that might be improved is very fine. What some see as blatant spam others see as an article needing clean up... this is particularly true when dealing with a company that doesn't have too much coverage---but does have enough to make wikipedia worthy.---I'm Spartacus! NO! I'm Spartacus! 00:50, 24 May 2009 (UTC)
I have to say that i got an utter dislike for spam pages and therefore tag those article for removal quite... vigorously. True, i won't tag article's that have a slight positive bias or contain a promotial word here or there - i just tag those as cleanup or peacock - but if an article is riddled with them they go to G11, no exception. I agree that some of these could be salvaged by cutting down the spam words, then rephrasing a few lines, and last by adding some non self published sources.. But i utterly refuse to do that for an article obviously meant as a promotion.
We have heaps of example articles regarding companies around and i have seen obvious CoI/company accounts create pages that aren't that bad at all. I had discussions with those account regarding improving those articles and some actually did help. But should we try and be lenient to accounts that are obviously here just to promote their own business? Should we rewrite their articles for them so they might actually have a wikipedia page? I say no. This might be my greatest flaw as a CSD patrol but i just get the shivers from articles that state lines comparable to "We are a leading company focused on customer friendliness and high quality. Our expert engineers can"... Yes, i could cut them down to a stub with some references but... Just no. Excirial (Contact me,Contribs) 07:28, 25 May 2009 (UTC)
And I see too many pages nomed for deletion per SPAM that are more marginal. If it uses terms such as "we" or "our" you are almost definitely dealing with a copyvio and can almost always delete it as such. But when it sends shivers up my spine when I look at some of the articles nominated for and deleted as blatant advertising where there might be a valid article worth keeping. It is, IMO, the worst defined criteria that we have. I personally don't touch it with a ten foot pole because I have so little respect for the way it is currently written.---I'm Spartacus! NO! I'm Spartacus! 08:24, 25 May 2009 (UTC)
I actually find this intriguing. I may dislike spam articles a lot, but yet you are stricter then me with article's written in the "We" form. I find that these article's are often not copyvio's but just an indication that a company is promoting themselves. Especially companies around India tend to write new article's in the "We" form - and they are often not copyvio's.
As for the way the criteria is defined: Yes, it are loose guidelines but there is no way to tighten these down without removing the G11 altogether - there are to many forms of spam. As for users overtaging with it (Or overtagging it in general): Just drop them a note. I didn't figure until my failed RFA that my tagging was way over the top of being acceptable. Hence, i thought i was doing a good job because 95% or so of the tagged article's were removed. In fact, i couldn't have been more mistaken and i didn't figure until a few people told and showed me that. Excirial (Contact me,Contribs) 09:11, 25 May 2009 (UTC)
  • I agree with a number of the points/suggestions, but not eliminating "no context". If the article could be improved with a bit of work then the tag has been misapplied. I remove the tag, do a bit of copy editing, and drop a friendly note to the tagger. I rarely see the same editor misapplying it again.--Fabrictramp | talk to me 16:22, 20 May 2009 (UTC)
    • As I'm Spartacus has pointed out below this is one of the categories with the highest error rate; So would you be happy with my amended proposal to make no context a prod criteria? ϢereSpielChequers 07:55, 22 May 2009 (UTC)
      • I wouldn't disagree with this, but i would prefer a non removable prod first. I would rather not have to AFD articles that contain a single line of text due to the formality that a prod can not be replaced on removal. Excirial (Contact me,Contribs) 08:57, 22 May 2009 (UTC)
  • I am interested in the idea of pages defaulting to userspace rather than namespace. It obviously requires tweaking, but it points out that many pages created are vandalism or short lived pages. Perhaps, instead of userspace pages that will persist without review (like free web storage space) indefinitely, an automatic expire of new-user pages that aren't otherwise approved. The problem with all of these ideas is that they overly complicate the system. I understand the RfC as a very straightforward and simple, albeit powerful, change. I think this idea is interesting, but I do not think it's especially relevant here. Shadowjams (talk) 07:49, 21 May 2009 (UTC)

I could support a prod that isn't removable by the user or creating time delay for most CSD categories. (COPYVIO/vandalism/Attack pages would have to be speedily deleted.) And we would have to break A7 into two categories. The first would be true vanity pages, which should be deleted speedily and Second those pages where significance might be establishable, but doesn't come close to meeting N/RS/V in the current status.---I'm Spartacus! NO! I'm Spartacus! 00:45, 24 May 2009 (UTC)

This is a much better idea than mine. The real ineffectiveness of the prod is that it can be removed by the creating user immediately after being placed. A time delay (not sure how realistic this is technically) would help as would policy changes to permit any user other than the creator to remove a prod. Shadowjams (talk) 05:25, 24 May 2009 (UTC)
It would work just fine - after all the current prod is already time delayed 7 days. All in all it would be little more then rewriting a policy to reflect this, and possibly adding an abuse filter to prevent this form of removal. Excirial (Contact me,Contribs) 07:03, 25 May 2009 (UTC)

Statement by M (talk · contribs)[edit]

Wikipedia works through consensus. Consensus is nebulous in definition, but its spirit is that anyone's views will be considered among the group and not just by an appointed individual, and that views are weighed by merit and not by count. This makes Wikipedia more inclusive, and it lets us catch all sorts of great points that a single person might miss. Sometimes, we suspend this inclusiveness and cooperation. An administrator might have to immediately block a spammer, and we usually don't need a team to insert a comma. But whenever possible, because we don't want to alienate newcomers, create hierarchies, or simply make careless mistakes, we try for consensus.

Unlike protection, which is neutral and intended to draw out discussion, speedy deletion is by definition non-neutral and intended to bypass discussion. Unlike blocking, deletion is seldom urgent. Deletion, even by a group, has a chilling effect upon and can upset well-meaning contributors - newcomers especially. It is even worse when just one person rules against the article you wrote. So we evaluate deletions in groups. This isn't a question of rules imposed on administrators, but one of the community's attitudes towards deletion. The exception to the usual caution towards deletion is that we permit speedy deletions, for those admins who want to take part in that, under 'mechanical, indisputable, and simple' criteria. "Use your best judgment" is not really such a criteria - nor is it even a criteria.

IAR presumably overrides anything in CSD regardless. Is there a need for the new clause here? "If you use good judgment, then you may delete without discussion, if you must" belongs in IAR. Even if administrators were perfect, there seems something amiss with explicitly endorsing admin bypassing of discussion in the CSD policy. Claim IAR against that rare case where an article must be killed quickly. I'll support your good judgment. But let's avoid claiming IAR where a process exists explicitly to handle article deletions that require time, community input, and courtesy to even misguided contributors. Just pass it along to XfD, and, keeping in mind community attitudes towards deletion (whatever those are), employ IAR when appropriate.

 M  05:25, 20 May 2009 (UTC)

Users who endorse this statement[edit]
  1. DGG (talk) 14:42, 20 May 2009 (UTC)
  2. Fabrictramp | talk to me 16:24, 20 May 2009 (UTC)
  3. Stifle (talk) 18:13, 20 May 2009 (UTC)
  4. --Elen of the Roads (talk) 18:32, 20 May 2009 (UTC)
  5. Ruslik (talk) 19:00, 20 May 2009 (UTC)
  6. Agree. If there is need for quick deletion, it is usually covered by G10 or G12 anyway and I cannot think of any case where there would be such need that does not fit current criteria. SoWhy 20:03, 20 May 2009 (UTC)
  7. Hobit (talk) 21:53, 20 May 2009 (UTC)
  8. MLauba (talk) 22:41, 20 May 2009 (UTC)
  9. ThaddeusB (talk) 23:53, 20 May 2009 (UTC)
  10. Almost precisely what I came here to say. --Tango (talk) 14:20, 21 May 2009 (UTC)
  11. Bearian (talk) 18:40, 21 May 2009 (UTC)
  12. Broadly. Davewild (talk) 21:41, 21 May 2009 (UTC)
  13. Endorse, with the emphasis that IAR should be rare. If a person is routinely calling upon IAR, then they need to re-evaluate their approach to CSD.---I'm Spartacus! NO! I'm Spartacus! 01:59, 22 May 2009 (UTC)
  14. My views as well. --R'n'B (call me Russ) 14:10, 22 May 2009 (UTC)
  15. IAR should be exceedingly rare, no one should use it repeatedly. Collect (talk) 14:00, 24 May 2009 (UTC)
  16. Agreed, assuming that the question is a rhetorical one. decltype (talk) 17:19, 24 May 2009 (UTC)
  17. --Aervanath (talk) 07:19, 25 May 2009 (UTC)
  18. Ynhockey (Talk) 09:28, 26 May 2009 (UTC)
  19. Mukadderat (talk) 05:12, 4 June 2009 (UTC)
  20. GRBerry 14:25, 4 June 2009 (UTC)
  21. Nicely done, M. - Dank (push to talk) 13:52, 11 June 2009 (UTC)
  22. Yes. TotientDragooned (talk) 18:25, 13 June 2009 (UTC)
  23. Yup.—S Marshall Talk/Cont 00:09, 19 June 2009 (UTC)
  24. Like Spartacus, with emphases that IAR should be rare. -- Ned Scott 05:56, 29 June 2009 (UTC)
  • I think the issue is not that IAR exists, or that IAR needs to be noted as a "clause" in CSD, but that the current incarnation of CSD effectively denies the existence of IAR. Happymelon 10:02, 20 May 2009 (UTC)
    The proposal up for discussion is to generalize and simplify the policy. My statement says, in part, that there is no need to add "feel free to ignore this, if you think it best" into every rule since this is already a pillar. But especially at CSD, where admins must use extra caution. And we certainly don't need to insert that as a means to reduce the specificity of the speedy-deletion criteria.  M  17:29, 20 May 2009 (UTC)
    I don't think we want to add a clause that says "feel free to ignore this." I think we want to remove or rewrite the parts of the policy that contradict that principle. And they are numerous. --causa sui talk 18:30, 20 May 2009 (UTC)
    Perhaps we have differing ideas of what it means to contradict IAR—if it's possible to contradict it (I don't think it is given its status as pillar). We need to be very clear about the kind of changes in your proposal: " that it does not imply that the policy is a firm rule", "Generalize the criteria...", "...further simplify the policy by reducing...". As I state more delicately above, I see this as an attempt to inject IAR into an unrelated policy so as to change that policy's meaning and purpose. This much is clear: whether these changes are necessary and justified is the subject of the majority of statements here, and I think they speak better than I can.  M  20:32, 20 May 2009 (UTC)
    Of course it's possible to contradict it, though nobody should pay attention to the policy if it did. An extreme example would be if the policy said "These are rules, and they are to be followed even if that prevents you from improving the encyclopedia." That's an extreme caricature, but it's not far off from what the policy actually does say. You're absolutely right that I am trying to change the meaning and purpose of the CSD policy, because its current meaning is incompatible with the basic ideas that make Wikipedia work. --causa sui talk 20:41, 20 May 2009 (UTC)
    IAR should be rare. Unfortunately, there are those who think IAR should be the norm, not the exception. I would be opposed to adding a new csd criteria for IAR because it would simply encourage more sloppy deletions... and people would essentially stop using any of the other criteria, because it would be easier to drop it into the IAR bucket.---I'm Spartacus! NO! I'm Spartacus! 02:01, 22 May 2009 (UTC)
  • As Ruslik points out, there is an arbcom case that deals explicitly with this entire issue here. It states that "Administrators are expected to use their deletion and undeletion abilities consistent with these policies. Administrators who wish to delete pages that lie outside the criteria for speedy deletion should usually list those pages at the appropriate deletion discussion forum, such as Articles for deletion or Miscellany for deletion, or apply a Proposed deletion tag.", and mentions that blatantly inappropriate content and WP:BIO violations may still be deleted. Passed 14:0. I hope that Ruslik or someone else will contribute the full quote as another statement in this RfC, so that it could receive proper attention.  M  21:41, 20 May 2009 (UTC)
    Sure, here you go:

The policy pages for Deletion policy, Undeletion policy, and Criteria for speedy deletion together provide policy and procedure for deletion and undeletion of pages, including pages in userspace. Administrators are expected to use their deletion and undeletion abilities consistent with these policies. Administrators who wish to delete pages that lie outside the criteria for speedy deletion should usually list those pages at the appropriate deletion discussion forum, such as Articles for deletion or Miscellany for deletion, or apply a Proposed deletion tag. This does not negate administrators' ability to delete blatantly inappropriate content even if it falls outside the formal CSD criteria, nor constrain application of our policy on biographies of living persons.

— Wikipedia:Requests_for_arbitration/MZMcBride
  • I would vastly prefer ArbCom's wording to the wording that is currently in the policy. --causa sui talk 21:49, 20 May 2009 (UTC)
    I disagree, and this is not a statement, but a comment, and therefore cannot get the attention and discussion that it needs. I hope that nobody is discouraged from bringing it up as a separate issue for discussion.  M  22:02, 20 May 2009 (UTC)
    Actually, i do not care if this is a mere comment, an arbcom statement or even something added to one of the cabal pages as a jest. When discussing an open issue it does not matter where words come from, what matters is how you interpret and utilize them.
That being said this wording is indeed preferable. Instead of locking down particular situations this wording implies that admins are bound to follow the policy and not entitled to make their own rules about removal; Yet it does allow for blatant violations to be removed. This is much more generic and much less patched up then the current wording of the CSD policy. M, as an obvious inclusionist, why would you disagree about this wording? It clearly sets a limit to what, and what can’t be done. The inclusion of blatant is clearly not a free warrant as it is mentioned right next to a statement that admins are supposed to follow the rules. If this wording would be made stricter then we might just as well archive all current removal policies and create a policy where either everything is notable or should be discussed. Excirial (Contact me,Contribs) 22:12, 20 May 2009 (UTC)
I think what he means is that we would need more eyes on this to get consensus to use Arbcom's wording instead of what is already there. I can accept that, but I'm nevertheless baffled about why he would cite this Arbcom quote to settle the debate in his favor and then object to adding it to the policy. --causa sui talk 22:28, 20 May 2009 (UTC)
What do you mean its not a statement? Its a principle from a decision on a case and was supported by 14 arbitrators, the other 2 recused from the case. If that's not a statement, what is? Mr.Z-man 00:57, 21 May 2009 (UTC)
If you think that's a statement, add it to the bottom of these headings ;)  M  02:22, 21 May 2009 (UTC)

Statement by Mr.Z-man (talk · contribs)[edit]

Users are given the admin tools because the community trusts their judgment and they have a history of dedication to the project.

  • We trust admins to use their discretion when protecting articles – the protection policy gives only general conditions in which to use protection.
  • We trust admins to use their discretion when blocking users – the blocking policy does not give specific criteria as to how disruptive a user must be to warrant a block or an exact scale for how long to make blocks.
  • There is no policy regarding granting or revoking rollback.

Protection can hamper article development just as much as deletion and blocking a user will surely annoy them more than deleting their article, yet no one proposes specific criteria for any admin tool except deletion. Admins have had no trouble applying these policies consistently even without strict criteria to follow.

There are no restrictions on IAR other than it must improve the project. If it improves the project you are free to do it. You may be asked to justify how it improved the project, but as long as you can do successfully do so, it was a correct usage of IAR. Anyone who claims IAR doesn't apply to deletion or IAR can never be used to bypass a discussion is misinformed about the purpose of and reasoning behind IAR.

For all the talk about some admins widely ignoring CSD and deleting articles "improperly" there are surprising few DRVs or "official" complaints being filed. This suggests either 1) the "issue" is grossly exaggerated or 2) people disagree with the deletion method, but agree with the results.

No other policy is designed to be followed exactly to the letter to the extent that the CSD policy is, not even close. People are using the CSD policy to change what the community considers an acceptable deletion reason, when it should be the other way around - what the community considers acceptable should determine the policy. Changes to the policy are affected with the agreement of only a handful of users; concentrating too much power in the hands of a few. Right now the policy attempts to assert itself as the only cases in which admins may delete a page without discussion outside of WP:PROD, attempting to eliminate entirely the idea of admin discretion in favor of a few specific, barely-subjective (and in some cases, completely mechanical) criteria.

Admins are responsible for their deletions whether policy justifies them or not. Policy justifying it does not make a bad deletion any less bad and policy does not excuse admins from using common sense. We give admins the tools because we trust their judgment, we should let them use both together.

-- Mr.Z-man 01:16, 20 May 2009 (UTC)

Users who endorse this statement[edit]
  1. This is an extremely clear and accurate description of the issues. Happymelon 09:59, 20 May 2009 (UTC)
  2. causa sui talk 13:39, 20 May 2009 (UTC)
  3. Excirial (Contact me,Contribs) 14:24, 20 May 2009 (UTC)
  4. Stifle (talk) 18:17, 20 May 2009 (UTC)
  5. --Elen of the Roads (talk) 18:58, 20 May 2009 (UTC)
  6. Shadowjams (talk) 07:52, 21 May 2009 (UTC)
  7. Bearian (talk) 18:41, 21 May 2009 (UTC)
  8. hmwithτ 11:33, 24 May 2009 (UTC)
  9. Carlossuarez46 (talk) 18:22, 15 June 2009 (UTC)
  10. A new name 2008 (talk) 01:17, 16 June 2009 (UTC)
  11. Cybercobra (talk) 23:39, 23 June 2009 (UTC)
  12. MacMedtalkstalk 18:20, 25 June 2009 (UTC)
  13. By Aristotle: Laws must be established, but they must not have authority insofar as they deviate from what is best, though they should certainly have authority everywhere else. CSD needs to be on the lines of the Blocking Policy or Protection Policy, not, so to speak, a section of the U.S. Code. —Animum (talk) 03:12, 27 June 2009 (UTC)
  • The only reason I do not challenge more is the need to keep good relations with my fellow admins. I trust most admins to use discretion most of the time. I trust nobody among us to use it right all the time. DGG (talk) 14:48, 20 May 2009 (UTC)
  • I find it's more productive to talk to admins one-on-one about dubious deletions than to haul them to DRV. (And I always appreciate the same courtesy myself).--Fabrictramp | talk to me 16:09, 20 May 2009 (UTC)
  • I suspect new editors back away rather than making a complaint, and some will leave the project altogether.--Elen of the Roads (talk) 18:41, 20 May 2009 (UTC)
    Surely. I would guess that for 99% of editors who have their first article speedy deleted, they don't come back because they've discovered that there are editorial controls on Wikipedia that prevent them from promoting themselves or the band they just formed, and therefore Wikipedia is useless to them. The percentage who created a legitimate article that was erroneously speedy deleted... I would say that has to be much lower than 1%. That doesn't excuse it when it does happen, and I think that if we are concerned about that, we should patrol Special:Log/delete for peer-review. --causa sui talk 18:46, 20 May 2009 (UTC)
The ones writing articles about their sister's band may grow into good editors one day.... Might take a while though. I do have a concern for those that set out to create a legitimate article, didn't make it, and got it speedied, but I wouldn't like to put a figure on how many that might be.--Elen of the Roads (talk) 18:57, 20 May 2009 (UTC)
Even if it were 1% it is 2500 a year. I know that every week at least 2 or 3 people in that situation go to the trouble of telling me by email that they are not coming back--and that's just me DGG (talk) 23:07, 20 May 2009 (UTC)
Sorry, is this for non-CSD deletions? Or just CSD deletions? Are these usually good-faith, or spammy contributors?  M  23:40, 20 May 2009 (UTC)
Some will undoubtedly just be spam, or pet rabbits. But some are setting out to write about some person/thing that is notable, but not celebrity status notable. Because of unfamiliarity with Wikipedia, they write a few sentences intending to come back with more, possibly not even realising that it's instantly published, and come back to find it speedied. I've had to rescue our town hall press officer once, when he did pretty much the same thing about something local that does pass the notability test, only you wouldn't have known that at the start. I think though, in fairness to your comment, that people in that position can also be affected by PROD and AfD, and take the bat home rather than improving the article.--Elen of the Roads (talk) 08:23, 22 May 2009 (UTC)
I'm referring more about the complaints by the people advocating for a stricter CSD policy. Mr.Z-man 19:44, 20 May 2009 (UTC)

You forgot options 3, 4, and 5. 3) The person whose article is a new user and is so turned off on the project that they leave or do not know how to go about appealing to DrV. 4) The person whose article was deleted is an experienced user, is still pissed off that their work in progress was deleted, but rewrites it and republishes the article. 5) That the only time DRV is necessary is when an article is deleted and the deleter and article creator can't agree to the resolution. Most of the time when an article is deleted, the deleter will avoid DrV by recreating the article. In other words, for every article taken to DrV, there are probably 10 articles that could have ended up there if the deleter didn't restore the article on their own. For every article restored on their own, there are probably 10 that should have been, but the article's creator didn't know about their options or left. Even if you say the numbers are one in five at both levels, there are still 25 articles improperly deleted for every case taken to DrV! CSD is the one area where there is the least amount of oversight and the people who suffer the most are most likely the newer users not your more experienced users.---I'm Spartacus! NO! I'm Spartacus! 02:11, 22 May 2009 (UTC)

Did you not read my comment directly above yours? I'm not talking about users filing DRVs for their own articles, I'm talking about the users (like yourself) who repeatedly complain about admins misusing CSD or ignoring CSD on a large scale. Yet the deletions themselves generally go uncontested... In any case, I prefer not to guess at numbers. If I don't know the actual numbers, I won't try to give them. Mr.Z-man 02:27, 22 May 2009 (UTC)
And taking articles to DRV would accomplish what? "Oh no, USER:X is bringing another case to DrV." I am not interested in that much drama.---I'm Spartacus! NO! I'm Spartacus! 02:58, 22 May 2009 (UTC)
Also, you seem to be making a mistake in your belief in what AFD/CSD/DRV is for. DRV is only necessary if you want to keep an article in the current shape, not to dispute a poor deletion. If the original author (or I) want to work on an article and get it to standards, then DRV is completely unnecessary. The Role of CSD is to delete articles that even if worked on would have no chance of surviving an AFD. AFD allows the writer a chance to salvage an article, to get it from the current shape to a version that would meeting the communities expectations. Articles that are deleted via CSD have no chance of being salvaged, an article that is sent to AFD gives a user a chance to salvage the article. It gives the user a chance to find out what is wrong with their article. It helps them learn how WP works. By giving them this opportunity, it gives them a sense that the process is fair. One of my first experiences at WP was defending an article. The AFD became the driving force behind my pushing that article to FA status---I decided then and there that I was going to get that article to FA because of the AFD. Had it been CSD'd, I might never have had that motivation. Had I not had the opportunity to plead my case, I might have left WP right then and there. AFD is not about deleting articles, but about finding out which articles have people who are willing to work on the article to get it into shape. Rarely is an article deleted where somebody is willing and eager to work on it! CSD short circuits that oportunity! That an article is not taken to DRV doesn't mean that the CSD was correct, it only means that *I* and others like me area not going to take the effort to rewrite the article. And if we were to rewrite it, we wouldn't bother with DRV!---I'm Spartacus! NO! I'm Spartacus! 03:28, 22 May 2009 (UTC)
See, now I read "*I* and others like me area not going to take the effort" to mean "I'm not going to DRV them because I know the deletion would be upheld" meaning that while you may disagree with the deletion, you know that other people won't. If you don't think it would be upheld, why not DRV it? If the article shouldn't be deleted, and you think the majority of people will agree with you, then there's nothing to lose by DRV'ing it. Surely a not-very-good article is better than no article at all?
And no, I'm not mistaken in what I think DRV is for. The DRV directions include "This page exists to correct closure errors in the deletion process and speedy deletions." DRV is most certainly designed to dispute a poor deletion. Mr.Z-man 04:45, 22 May 2009 (UTC)
Absolutely not what I am saying. You seem to fail to understand the importance of sending an article through AFD as compared to CSD. Just because an article in its current shape would certainly fail an AFD does not make it case for IAR. Sending an article through AFD serves an important role, and that role is to allow others (particularly newer users) the opportunity to defend their article and learn Wiki policy. An article can be imporperly deleted that would not be overturned at DRV---heck, I've declined CSD's and then voted DELETE at AFD indicating that the article might be snow closed. Sending an article 3 months---even a week---later that was improperly CSD'd serves no purpose because the system already failed. But as you apparently don't appreciate the importance the various roles serve.---I'm Spartacus! NO! I'm Spartacus! 00:31, 24 May 2009 (UTC)
Yes, I care more about the content than the importance of the processes. Mr.Z-man 00:55, 24 May 2009 (UTC)
And thereby, IMO, chase away potentially quality users and delete articles that could be salvaged. This is the attitude at CSD that makes CSD one of the least respected areas of the community AND can cause the most long term damage to the project.---I'm Spartacus! NO! I'm Spartacus! 01:00, 24 May 2009 (UTC)

Statement by Excirial (talk · contribs)[edit]

Ever since wikipedia has been founded Wikipedia has been a growing community, which has probably managed to assemble the largest collection of information humanity has ever created, even surpassing the great Library of Alexandria. We have managed to do so by allowing everyone to add and edit article's, therefore creating a huge community with incredibly diverse knowledge and specializations of even the most obscure topics.

However, during the years we have also come to the conclusion that while such an open model is the core for wikipedia's success, it also carries some risks along with it. Such a great collection of information sometimes seems to imply that anything can be added to Wikipedia regardless the topic, regardless the importance of the subject. In order to keep wikipedia clean from unwanted pages we have the means to remove them trough discussion, which is most times handled at the articles for deletion section. Yet due to the huge influx of articles this section would soon be swamped if we needed to talk about every deletions, regardless if a page had any chance that it could actually stay.

In order to prevent discussions that can only end in one conclusion the CSD policy was created, which allows article removals without the need for a removal discussion prior to the actual removal. This policy bases itself on previous removal discussions that indicate there is consensus for removal for certain pages.

In practice we learned the risks of incorrect deleting of articles without discussion and therefore we buckled down the CSD guideline to such an extend we assume it allows for virtually no mistakes. Yet while buckling it down we forgot that Wikipedia is an encyclopedia about everything - and man cannot make specific rules about every single topic. This is painfully visible in the A7 categorie which allows for removal or persons, organisations and web content that do not seem notable. This particular lockdown creates the rediculous situation in which persons can be removed without discussion while pets, furniture and other subjects that would virtually never be notable have to go trough the discussion based deletion routine. If we trust administrators to judge over topics that have a high chance of being notable, why do we lock the policy down in a way that they are not allowed to judge over articles that are likely to be non notable?

My suggestion is that we stop artificially limiting the categories which can receive a notability CSD, but instead focus on the quality of the tags placed. In practice this will lead to fairly little change: Articles that clearly have no chance of surviving an AFD but won't practically fall under the A7 policy are shoved under other categories of the deletion policy, or they are speedy removed trough the application of The IAR guideline. This is simply a necessarily to prevent both removal discussion that can only lead to one conclusion, as well as to prevent literally drowning AFD volunteers in articles. The original guideline was created in 2004 with one of the reasons being that we could otherwise not handle the flood of pages for discussion on AFD. Seeing this was 5 years ago, with wikipedia's article size being 7x smaller then what we have now the amount of pages created each day is an order of magnitude larger as well.

If we fear that administrators cannot be trusted handle an increased scope of the A7 policy correctly and fairly then we should ask ourself why we even appoint administrators; Giving them trust to use certain tools before hermetically buckling them down is of no use to anyone. If we do not trust an admin to judge over a page this is probally better solved by splitting the task between two admins, with one deleting a page and the second reviewing that particular removal, rather then trying to apply some artificial limit of trust.

Excirial (Contact me,Contribs) 09:10, 20 May 2009 (UTC)

Users who endorse this statement[edit]
  1. causa sui talk 13:45, 20 May 2009 (UTC)
  2. Per my comment below. Mr.Z-man 15:13, 20 May 2009 (UTC)
  3. Stifle (talk) 14:39, 21 May 2009 (UTC)
  4. hmwithτ 11:34, 24 May 2009 (UTC)
  • The reason for the limitation is that these have proven the only categories that can be safely judged by two people and that occur with any frequency. I'll expand on that in my own statement.DGG (talk) 15:00, 20 May 2009 (UTC)
  • PROD is basically useless for a new article; the author will almost certainly remove the tag, so the only real options are AFD and speedy. A rather significant amount of AFDs close with unanimous decisions to delete; I think AFD would be a much more effective process if people didn't spend so much time on things that didn't need a discussion. Mr.Z-man 15:13, 20 May 2009 (UTC)
Prod works--there are dozens of expired prods deleted each day. Most people recognize when something is hopeless. DGG (talk) 15:18, 20 May 2009 (UTC)
Prod works for older articles that haven't been edited in a few weeks. For brand new articles, its almost a waste of time. Mr.Z-man 15:41, 20 May 2009 (UTC)
That sounds more like an argument for a prod defender bot to stop authors removing prods from articles they've created, than an argument against prods per se. ϢereSpielChequers 16:29, 20 May 2009 (UTC)
An issue with prod is that anyone is allowed to remove them - even if you are the writer of a hopeless page. This can be a good thing as it allowed for people to improve the page and afterwards delist the page so we no longer need to judge it. However, in a lot (Maybe a majority) of the cases the author removes a prod without even changing the article, forcing it to be forwarded to AFD since a prod cannot be relisted. In fact a prod template requires to be monitored in its 7 day lifetime to see if an article is improved; This means that the person placing the prod will spend more time then it takes to make an AFD. Also, it will eliminate the advantage of PROD as means of non controversial removal of possibly improvable articles. Excirial (Contact me,Contribs) 18:14, 20 May 2009 (UTC)
That bot's beginning to sound more appealing by the minute :) I wonder though what kind of administrative burden it would create if all PRODs had to be kept in place and examined by an admin at the end of the seven days.Elen of the Roads (talk) 18:29, 20 May 2009 (UTC)
Isn't this already the case? The only difference would be that there would be more of them, since fewer PRODs would be (illegitimately) removed from articles by their creators. But that's a good thing, isn't it? --causa sui talk 18:33, 20 May 2009 (UTC)
Well, I would have said it was a good thing. But it isn't what happens now - PROD's are deleted by content authors without making any changes to the article. The PRODder can't put it back, it doesn't go to an admin, nothing happens. This is against the spirit, but not against the rules. You could argue that in such cases the content author thinks that the article is fine as is, and it is correct for it to go to AfD if anyone wants to delete it, and I'm sure that's sometimes the case. I'm equally sure other times that the thoughts of the content author are less exalted.Elen of the Roads (talk) 18:48, 20 May 2009 (UTC)
How about: We put an equivalent of a "hangon" template for author-contested PRODs. These won't be automatically deleted, but manually reviewed before deleting them. The bot prevents the author from removing the PROD outright, but the author can request that it be reviewed by another editor before it gets deleted. I'm just brainstorming here. :-) --causa sui talk 18:53, 20 May 2009 (UTC)
Tsk! Solutioning before we've fully mapped the problem :) That would be a useable solution I think - don't remove the PROD but add a "contested" template. Elen of the Roads (talk) 19:28, 20 May 2009 (UTC)
It's just not true that "we buckled down the CSD guideline to such an extend we assume it allows for virtually no mistakes." My experience as an admin patrolling CSD is that there is plenty of room for mistakes, especially (but not exclusively) A1 and A7, and other users' comments here seem to corroborate that. If you think A7 is too restrictive, then propose a change to that guideline. Just suggesting to admins to delete whatever they think is consistent with the "spirit" of CSD will make the situation more, not less, prone to error. --R'n'B (call me Russ) 14:27, 22 May 2009 (UTC)

Statement by DGG[edit]

Some policies are more prescriptive than others. Some places one can use IAR more freely than others. Many things we do not use IAR on, such as adding copyright material beyond the NFCC, even if it would improve the encyclopedia. IAR could be used for speedy if a previously unforeseen case should arise that needed it. IAR is fundamental., but not as a first choice, rather if nothing else applies that would meet the need. In a sense it is used here, in practice by expanding a little on the definition of vandalism. We do that anyway in many places because of the urgent need to remove vandalism and the difficulty of defining all types of it, so it makes sense.. On the other hand, a decision at AfD can use IAR if it really needs to, because there is community involvement. We frequently use IAR to correct errors at Deletion Review, because correcting errors is a place where it's likely to be needed--I have invoked it there myself.

I have never seen an article that needed IAR otherwise in 2.5 years of patrolling CSD. The only things that need to be removed immediately are vandalism and BLP violations, and they always are. Clear copyright violation needs to be removed promptly, and it is. The other criteria are at speedy because it has proven relatively safe for two people to delete them. Things like books and products need a period of exposure at prod or afd. Those are areas where a poor but recoverable article may not be recognized if one does not know the general field. Many other things mentioned, like pets, occur very rarely. And if there is agreement enough, they can be added. We did add a criterion recently, A9 for records by deleted performers, because its unambiguous.

I do not see this as descriptive vs. prescriptive, because what we should be describing is what we think to be the right way to do things. We cannot describe the full range of what we do and leave it as that, because there are a small number of admins who delete on the basis of what they personally think ought not to be in the encyclopedia. The great majority do not, and such deletions are always reversed if taken to Deletion review. We do wrong things sometimes, and sometimes they stand because they are not worth the difficulty of challenging them, & the need to keep good relations between admins by not challenging each others actions unless they are really outrageous. . There is no necessary conflict between prescriptive and descriptive, because we should not and do not write a prescriptive rule that does not match what is generally perceived as good practice. A descriptive rule does not say the wrong way is right because a few people do it. It's like a law saying merely "most people do not shoplift".

In some cases I favor explicitly descriptive rules. I joined with MASEM in trying to get WP:PLOT to merely describe the range of what we do, and say it was not being prescriptive, because there was no agreement about what we ought to do, but that approach was rejected there. I continue to think it was appropriate, because inclusion guidelines are inherently flexible, and there is no likelihood of abuse because each case would have a community decision. The rules for what admins can do are another matter: the possibility of abuse is there--and the fact of it is unfortunately sometimes there also. The only way to prevent it is to be restrictive. I prefer not to word it descriptive/prescriptive -- I am using these terms only because others were. The clearer wording is restrictive/permissive, or fixed/flexible. I continue to think, and i believe it will be found that Wikipedians in general think, that there should be restrictive limits on what admins can do. In the nature of human behavior, they will go over the limits anyway, so the only way to keep it from getting abusive is to insist on it being narrow limits. Discretion is of course inevitable, as nothing can be applied unthinkingly. The way to keep discretion reasonable is for the policy to limit it. To actually say, this is what you should do , but you may do as you please if you really want to, is not going to produce good results here. There is a fundamental dilemma involved at CSD: our is that the bias should be towards improving and keeping articles, yet about a third of the articles submitted are hopelessly unimprovable. We absolutely need the procedure, but its dangerous.

Users who endorse this statement[edit]
  1. Agreed. We don't need to make IAR a fundamental part and process of CSD, since it's already that for WP in general.  M 
  2. Ruslik (talk) 19:13, 20 May 2009 (UTC)
  3. Amen to that. SoWhy 20:01, 20 May 2009 (UTC)
  4. Hobit (talk) 21:55, 20 May 2009 (UTC)
  5. Spot on. MLauba (talk) 22:43, 20 May 2009 (UTC)
  6. Agreed. I see "hopeless" products, books, etc. on PROD everyday that are, in fact, notable just very poorly written. I imagine many of these would be speedied if admins were not restricted. --ThaddeusB (talk) 23:52, 20 May 2009 (UTC)
  7. Yep. Like it. Ale_Jrbtalk 08:46, 21 May 2009 (UTC)
  8. Bearian (talk) 18:42, 21 May 2009 (UTC)
  9. Very well said. Davewild (talk) 21:44, 21 May 2009 (UTC)
  10. A7 and A11 could be revisited, but the problem is sloppy deleters.---I'm Spartacus! NO! I'm Spartacus! 02:18, 22 May 2009 (UTC)
  11. Take out 'no context' as suggested elsewhere (it's the one that's really too woolly) and I'm with you all the way Elen of the Roads (talk) 08:27, 22 May 2009 (UTC)
  12. There are plenty of issues about the current CSD criteria; they are too restrictive in some areas and too permissive in others, but just encouraging admins to ignore them is not the solution. --R'n'B (call me Russ) 14:13, 22 May 2009 (UTC)
  13. Concur. --Jack-A-Roe (talk) 04:04, 23 May 2009 (UTC)
  14. --ragesoss (talk) 05:49, 23 May 2009 (UTC)
  15. hmwithτ 11:36, 24 May 2009 (UTC)
  16. decltype (talk) 17:22, 24 May 2009 (UTC)
  17. --Aervanath (talk) 07:20, 25 May 2009 (UTC)
  18. lifebaka++ 18:00, 1 June 2009 (UTC)!
  19. Ched :  ?  03:19, 3 June 2009 (UTC)
  20. GRBerry 14:27, 4 June 2009 (UTC)
  21. Dekimasuよ! 13:42, 11 June 2009 (UTC)
  22. TotientDragooned (talk) 18:27, 13 June 2009 (UTC)
  23. Yeah.—S Marshall Talk/Cont 00:11, 19 June 2009 (UTC)
  24. Yep. - Peregrine Fisher (talk) (contribs) 23:59, 19 June 2009 (UTC)
  25. Cybercobra (talk) 23:44, 23 June 2009 (UTC)
  26. -- Ned Scott 05:54, 29 June 2009 (UTC)
  • I would point out that IAR could not be used to override NFC. NFC is basically an extension of a foundation resolution and partly based on copyright law, neither of which can be ignored with IAR. Even if they could, doing so would add legal risk to the project and compromise one of our core values, neither of which could be justified by a little extra decoration in an article. Mr.Z-man 00:53, 21 May 2009 (UTC)

Statement by Dycedarg (talk · contribs)[edit]

There is a definite difference between the areas in which admins exercise full discretion, as outlined by User:Z-man, and deletion. It's not a matter of the potential damage; bad blocks and page protections can be as disruptive and problematic as bad deletions. Where deletion differs from these other two examples is that it requires actual factual knowledge that may not be readily available. Blocking someone or not blocking someone requires knowledge of their actions and the backgrounds of the dispute that predicated them; this knowledge is generally readily available and as likely to be held by any given admin as any other admin. Protection requires a similar quantity of knowledge, specifically about what the activities on a page have been and whether they are likely to be resolved through blocking or lesser actions before full or semi-protection is employed.

Deletion however, requires that you know for certain that a given article does not fulfill inclusion requirements. This could be easy, in the case of someone's cat or the more obvious cases of attack pages and vandalism. But the determination of whether or not a large number of subjects should be included in Wikipedia is not particularly easy to make. Something that is obviously not notable to one person could be massively notable to another because of information that the first person simply did not have. I don't want admins to have the capability to go deleting anything they want based off of a failed Google test and their own personal certainty that a subject is utterly not notable. There are quite simply many absolutely notable subjects which are esoteric and unusual enough that a random layman is not qualified to make a judgment relating to their notability.

Do I think most admins would go deleting things relating to subjects of which they are wholly ignorant given the ability? I'm not at all sure. I'd like to think that most would not. But I can not be certain, and I don't want to have to be. I agree that CSD may be too constrictive at the moment. But I absolutely disagree with any alteration to the language that would allow admins to determine notability in cases where it is asserted based on their own personal judgment.

--Dycedarg ж 16:08, 20 May 2009 (UTC)

Users who endorse this statement[edit]
  1. Deletion based on the lack of notability requires general knowledge in the field in question. However, most admins (including me) are not experts in everything, and many do not know the limits of their expertize. I think, this statement is closest to my position on the issue under discussion here. Ruslik (talk) 19:23, 20 May 2009 (UTC)
  2. Given that 'judgment' appears to be the proposed criteria, this seems consistent with the sort of mistakes admins who misunderstand CSD might make.  M  22:06, 20 May 2009 (UTC)
  3. Agreed. I see "hopeless" products, books, etc. on PROD everyday that are, in fact, notable but are very poorly written. I imagine many of these would be speedied if admins were not restricted because they look utterly unnotable at first glance. --ThaddeusB (talk) 23:56, 20 May 2009 (UTC)
  4. Don't completely agree with everything, but you raise an excellent point about the difference between CSD and protecting/blocking. Another difference is that there are better controls over protecting/blocking than there are on CSD. If a person is blocked or a page is protected, the user is readily notified on how to get unblocked or unprotected. A poor block/protection, will more likely be overturned, because others will know about it. A poor deletion is probably never noted and the person whose article is deleted may not know how to go about appealing the trangression.---I'm Spartacus! NO! I'm Spartacus! 02:22, 22 May 2009 (UTC)
  5. Good outlining of why deletion is different. decltype (talk) 17:24, 24 May 2009 (UTC)
  6. GRBerry 14:28, 4 June 2009 (UTC)
  7. TotientDragooned (talk) 18:29, 13 June 2009 (UTC)
  • It has not been proposed (nor should it be) that administrators should have discretion to speedy delete articles based on notability, since that is not even in the criteria as they currently stand: the criteria is "asserts significance" which is much stronger, and nearly impossible to get wrong. The proposal does not lead to all deletion being in the hands of the first administrator that passes by: the proposal leads to administrators being able to delete articles about someone's screwdriver collection, despite there being no CSD criterion for "C163.3 insignificant screwdrivers", without getting harassed about the rules. --causa sui talk 16:16, 20 May 2009 (UTC)
    From what I can tell this RFC looks to change the policy. It wants to make it less restrictive and less based on technicalities, with the view of allowing admins to do what's best for the encyclopedia without being restricted by picky rules of no consequence. Well, I had seen no indication of what technicalities exactly are being thrown to the wayside. I've seen people state that admins should be allowed to delete obviously non-notable subjects via CSD, and such a person could potentially view the difference between an assertion of notability and actual notabiliy to be such a technicality that should be gotten rid of. I am merely objecting to that possibility ahead of time.--Dycedarg ж 16:31, 20 May 2009 (UTC)
    Right. What, exactly, will be "thrown by the wayside", to use your language, is a separate question. First, we have to agree that the policy should be rewritten to make clear that it is not prescriptive, and that some kind of simplification is necessary. What that would entail would be the result of further discussion. I am rather confident that nobody will be proposing adding "notability" as a criteria when that happens. If anything, I would expect several criteria to be removed. --causa sui talk 16:47, 20 May 2009 (UTC)
  • Your are right - the policy change aims to make the CSD policy less restrictive and less based upon exact wording of the policy. However, "Less restrictive" should not be interpreted as a right to delete articles at will because an administrator feels like doing so. The change will merely mean that an administrator is not bound to be restricted by wording that hampers the efficient functioning of their removal duties. As I stated before I actually had to take an article about a pet cat to AFD once because the speedy was denied, exactly as the rules were written, while at the same time tagging non notable persons for speedy deletion.
We receive little complains about incorrect removal of these and that should stay that way after a policy change. Yet, an editors and admins should be allowed to use common sense on a policy as broad as this one. Not being able to tag cats, dogs, screwdrivers, furniture and perhaps even the most severe violations of WP:NOT is illogical if we are certain they won't survive. Or as the [{WP:BOLD]] policy states so nicely: "Be bold, but be carefull" Excirial (Contact me,Contribs) 18:31, 20 May 2009 (UTC)
  • Again, we only give the admin tools to users with demonstrated good judgment and dedication to the project; they aren't random users. There are many unusual and esoteric subjects, but a) the vast majority of articles are not about these types of subjects and b) no one has yet presented an argument as to why admins will constantly screw up on these types of articles. Why do you think admins will totally abandon common sense with these articles? Giving admins the ability to delete things based on their own discretion outside of strictly defined criteria does not mean that speedy deletion would be the only deletion option for admins. Mr.Z-man 17:54, 20 May 2009 (UTC)
    • The problem is in the other direction--we need to rewrite the rules more clearly so there will be a sharper distinction between what you can and cannot delete--less discretion, not more. If it takes discretion at all, & its not a matter of imminent damage, it's a matter for the community. There's not a single admin here that I trust to be right all the time, most certainly including myself. As arb com has said many times, we're not expected to be perfect. Therefore we must arrange it that our imperfections do as little harm as possible. DGG (talk) 23:13, 20 May 2009 (UTC)

Statement by Kww (talk · contribs)[edit]

The CSD policy is one that must be kept tight and specific. We are talking deleting articles essentially without discussion or review, because the vast majority of editors can't view an article deleted by CSD to see whether it truly merited deletion. It makes sense that for two people to do that, they must be following clear, bright-line rules about completely unacceptable content, and be able to be held accountable for failure to do so. Certainly IAR comes into play, but IAR is a policy best applied sparingly. An admin that deleted articles that did not come under speedy deletion criteria on a regular basis would be abusing his trust: it's hard to envision that these articles occur so frequently that the trip through AFD would clog the system or be otherwise damaging.—Kww(talk) 16:11, 20 May 2009 (UTC)

Users who endorse this statement[edit]

  1. I agree. I've said something similar below. Shorten the prescriptive, take out all the janitorial function (dead redirects etc). Pages that have ceased to have a function are not a risk to the project, there is no need to have them described in huge detail. Clear direction about what is so dangerous to the project or such a nuisance that it must go out the door straight away.--Elen of the Roads (talk) 16:19, 20 May 2009 (UTC)
  2. Fabrictramp | talk to me 16:39, 20 May 2009 (UTC)
  3.  M  17:30, 20 May 2009 (UTC)
  4. Ruslik (talk) 19:28, 20 May 2009 (UTC)
  5. SoWhy 20:09, 20 May 2009 (UTC)
  6. Exactly. Hobit (talk) 21:56, 20 May 2009 (UTC)
  7. Agree & per Elen. MLauba (talk) 22:45, 20 May 2009 (UTC)
  8. Administrators unilaterally deleting articles is enough of a problem as it is; loosening the CSD criteria would only encourage such disruption.  Skomorokh  08:27, 21 May 2009 (UTC)
  9. Yep. Ale_Jrbtalk 08:48, 21 May 2009 (UTC)
  10. Hear, hear. Colonel Warden (talk) 14:20, 21 May 2009 (UTC)
  11. Bearian (talk) 18:44, 21 May 2009 (UTC)
  12. Yes. Davewild (talk) 21:46, 21 May 2009 (UTC)
  13. Short sweet simple... probably the best statement that anybody will make on the RfC!---I'm Spartacus! NO! I'm Spartacus! 02:24, 22 May 2009 (UTC)
  14. First choice. Regulate fishing, not fish. NVO (talk) 07:42, 22 May 2009 (UTC)
  15. Well put. --R'n'B (call me Russ) 14:16, 22 May 2009 (UTC)
  16. Enough stuff appears to get zapped at present there appears to be little reason to try to shoot down more.Geni 01:39, 23 May 2009 (UTC)
  17. Concise and to the point. decltype (talk) 17:29, 24 May 2009 (UTC)
  18. --Aervanath (talk) 07:22, 25 May 2009 (UTC)
  19. Juliancolton | Talk 19:23, 25 May 2009 (UTC)
  20. Ynhockey (Talk) 09:33, 26 May 2009 (UTC)
  21. Sjakkalle (Check!) 11:44, 31 May 2009 (UTC)
  22. Mukadderat (talk) 05:15, 4 June 2009 (UTC)
  23. GRBerry 14:29, 4 June 2009 (UTC)
  24. - Dank (push to talk) 14:17, 11 June 2009 (UTC)
  25. TotientDragooned (talk) 18:30, 13 June 2009 (UTC)
  26. A new name 2008 (talk) 01:33, 16 June 2009 (UTC)
  27. S Marshall Talk/Cont 00:12, 19 June 2009 (UTC)
  28. Totally. -- Ned Scott 05:58, 29 June 2009 (UTC)
  29. JamieS93 21:38, 30 June 2009 (UTC)


  • Ironically, I'm tempted to endorse your statement, because we agree much more than you might think. I don't think it's necessary to have the policy be read as prescriptive in order to hold people accountable to their actions. What counts is that the standard by which they are held accountable: what's wrong with bad edits is that they harmed the project, not that it was against the rules, as if the rules were ends in themselves. In that sense, you can have a policy that is as specific as you want, without creating the illusion that Wikipedia has firm rules. I'm not clear from your statement what you think about that particular issue. Can you clarify? --causa sui talk 16:13, 20 May 2009 (UTC)
  • Admins are just as accountable for their deletions whether CSD authorizes them or not. Policy does not excuse users from applying common sense and good judgment. Mr.Z-man 17:58, 20 May 2009 (UTC)
    • Yes, but unlike other decisions, with CSD users have almost no way of finding out whether they did really use their best judgment and follow policy. Regards SoWhy 20:08, 20 May 2009 (UTC)
      • So therefore we should allow them to judge people and companies and buckle down the rest to prevent this? That's what we would call damage control - trying to heal an infected wound by placing a patch over it so we at least won't see it as much. I can understand your point that there are little ways to check CSD behaviour as regular users cannot acces deleted pages, and admins need no form of discussion for removal; Hence, it is like a mysterious dark alley where we cannot go to see what’s there. However if we deem the lack of checkability/trust a problem then that is what we should solve, and not focus on limiting the amount of darkness that is in the alley. We could, for example, use a prod like system where 1 admin deletes a page and another admin reviews to prevent bad quality removals. While this might work we could also ask ourselves why we make people an administrator if we cannot trust them with this – WP:RFA is bad enough to survive in one piece as you will be completely reviewed before allowing usage of the mighty delete button… Unless you suggest they got psychological traumatize from the process and can’t function properly anymore… (Just a jest :P) Excirial (Contact me,Contribs) 21:20, 20 May 2009 (UTC)
  • Kww's point overplays the effectiveness of the current deletion criteria, and underestimates the deference most admins have for current wiki policy. Affirmative CSD enforcement has always been discretionary, which is something often overlooked. Any editor (other than the OP) can deny a CSD, but only an on-board admin can approve it.
This proposal goes a little ways towards fixing this imbalance. Ultimately it only asks that admins exercise the discretion in CSD criteria they're given in nearly every other area of their control.
There's no better example than the number of editors that have been roasted at RfA for getting the technical criteria wrong, without any suggestion of mal-intent. Justifications for not biting newbies, or not having unreviewed deletions are incredibly important. But we need to remember the other side of the deletion debate.
Delaying deletion provides free google based advertising for a very profitable period, adds to the bloat of the encyclopedia which decreases the focus of editors on valid articles, increases issues of synchronization (contradictions between articles), and decreases the manageability of the project. Few of us want an overly managed project, however loose article criteria force an overly managed project while simultaneously decreasing the effectiveness of whatever management we can come up with. Shadowjams (talk) 08:11, 21 May 2009 (UTC)
  • I see all three proposals as loosening the criteria. Currently, both the criteria themselves and the latitude of their interpretation are narrow and strict, and for very good reason. Making CSD less of a firm rule (proposal 1), generalising the criteria to allow more leeway for judging administrators (proposal 2) and simplifying or merging criteria in order to make their language less technical (proposal 3) would all clearly widen the scope of which articles can be speedily deleted. This would essentially make speedy deletions a matter of local administrator whim rather than global community consensus. Patrollers and administrators, from what I have observed, have shown themselves as a whole not competent enough to exercise this degree of free reign. Regards,  Skomorokh  11:02, 21 May 2009 (UTC)
  • On the contrary, it would make deletion more difficult since it would make administrators responsible for their actions and (potentially) obligated to explain every one of their deletions in terms of why it helped the project, rather than why it was consistent with the already highly generalized policy. --causa sui talk 11:15, 21 May 2009 (UTC)
  • I start to feel that this is useless - no matter how many times i state this is not the reason for changing the policy, yet it seems people want to believe their own interpretations. One more try: I have already stated before that CSD is used to loosely at times, due to which articles that are not completely hopeless get removed trough it without due discussion. This is a real concern as this will scare away editors, and is completely against the spirit of CSD. Personally my opinion is that the A7 policy is the main culprit in this as notability is really easy to use as a removal reason.
Yet what did we do to solve this? Instead of trying to set a hard limit based on the article quality we set a limit on article category. In other words: We try to tell ourselves that by limiting the amount of CSD tags to, say, 70% of the articles we prevent all trouble with these tags. We fool ourselves thinking that we can offset the questionable (As in: Maybe we should AFD for consensus) removals within that 70% by being extremely strict on the matter which articles can be removed. We will cry murder and damnation if a tag does not fall exactly within a category, yet we seem to be ignorant of the fact that the 70% of the articles that are allowed to be tagged have a percentage of to lenient tags themselves. Thus: We deem it necessary to discuss completely worthless articles about pets trough AFD, yet we seem to have a consensus that articles about persons can be deleted trough CSD. This is exactly what CSD currently states. Truely, Would we not rather discuss those to lenient tags concerning articles about persons, and just remove articles about certainly non notable things?
This is why i prefer MZMcBride wording to be utilized in the policy. It clearly states that CSD should be taken strictly, with only a rare exception if the situation calls for it. We should either have strict wording in a policy and a broad policy, or lenient wording and a very specific scope that still does not trade in quality. Yet we currently seem to think we can remedy the matter of "CSD already removing to many articles" by simply allowing this to happen on less categories. When in doubt, AFD or PROD it. When absolutely certain, CSD it. But limit on quality, not category. Excirial (Contact me,Contribs) 12:07, 21 May 2009 (UTC)
The comments by Kww here are very wise (and I have rarely agreed with him!). Several occasions where an administrator speedy deletes articles without regard to the CSD policy have created a lot of ill will, and a lot of time-wasting bureaucracy while the issue of undeleting the material is sorted out. The fairly well-defined boundaries for what is a speedy deletion candidate prevents (or at least sharply reduces) differing standards among admins from influencing whether something should be speedy deleted. With streamlined processes such as PROD in existence, and with the vast majority of AFD discussions being closed in a timely manner with little backlog, there should usually be no need to go beyond these criteria. Sjakkalle (Check!) 11:44, 31 May 2009 (UTC)

Statement by Elen of the Roads[edit]

I think there are two issues with CSD, as it currently stands, that are probably causing admins to feel the way they do.

First - it is way too long. It repeats itself (look at Patent Nonsense and see how many of its examples reappear as classes of their own) and it's too full of instructions. Second - it's a hotchpotch. Classes that are potentially dangerous to the project (copyvio, attack pages) or have real nuisance value (gibberish, hoaxes, vandals), are mixed in with classes that probably represent someone unfamiliar with the rules (no context, no content, no notability, spam) and classes that are purely maintenance (all the various left over pages)

So admins are getting the same set of detailed, prescriptive instructions for everything. If however one shortened the "this is policy" "this is stuff that must be dealt with urgently" section to the dangerous and nuisance classes - admins 'must' should delete copyvio, attack pages, outrageous BLP with no sources, gibberish, vandalism and blatant hoaxes, offerings from banned users as a priority - then that would give a clear signal to everyone as to what will be deleted with vigour, while at the same time making the document simpler.

Creating a "maintenance" or "janitor function" category for all the orphan pages, dead redirects, author requests, transwikied items etc etc. would also seem appropriate. This is where it should be reasonable to trust the admins with just a general instruction "any page that has ceased to have a function, bearing in mind that some pages are retained because of the page history."

Having separated out these clear categories, the opportunity is then there to have a further discussion on the category of classes that might reflect new editors, mistakes, people unsure of the rules etc, which perhaps moves on from merely hustling them out of sight, and towards identifying whether this is someone who needs support or an indefinite block.Elen of the Roads (talk) 16:16, 20 May 2009 (UTC)

Users who endorse this statement[edit]
  1. causa sui talk 18:30, 20 May 2009 (UTC)
  2. Lankiveil (speak to me) 05:55, 21 May 2009 (UTC). Surprisingly (to me), I actually agree with this; the existing CSD is a bit of a hodgepodge and doesn't draw a line between stuff that has to go straight away (G10, G11), stuff that probably should be culled quickly in the interests of the credibility of the project (G12, G1, A3), and stuff that can probably wait around for a few hours at least to see if it can be salvaged (A7, A1). It seems logically that we should deal with these different sorts of problems in different ways.
  3. Mr.Z-man 19:15, 21 May 2009 (UTC)
  4. Good idea, although I suspect that refactoring CSD will be an administrative mess. NVO (talk) 07:46, 22 May 2009 (UTC)
  5. In general, I agree with this, but not with every detail. I think the only deletions that are urgent are copyvios and potential defamation (BLP, attack pages, etc.). Other than that, there's no deadline; no one is going to be harmed by stumbling across an article about a user's horse. Those should be deleted to maintain the quality of the encyclopedia, but they are still a different class of problem than the harmful ones. And we definitely need to find a way to be more welcoming to new users who just don't yet understand what makes an article acceptable. --R'n'B (call me Russ) 14:21, 22 May 2009 (UTC)
  • Do you think that it would get the same work done to simply convey the urgency of deleting such articles, especially BLP problems (attack pages, etc), rather than making it into a commandment? I would think that someone who is here to improve the encyclopedia would go about immediately deleting these pages once he or she had had it clearly explained that it is of the highest priority to remove this content from the encyclopedia immediately, and here's why. Do you think that could achieve the same effect in practice? --causa sui talk 16:26, 20 May 2009 (UTC)
    Actually, I'm sure it would. It doesn't need to say "this is policy". It could just as well say "this content is a huge problem. Shifting it is of the highest priortiy" I think the key is splitting the risk items out from all the others, because the janitorials will wait, and the 'new editor' ones should allow the Admin some leeway, and encourage more dialogue with the author.Elen of the Roads (talk) 16:45, 20 May 2009 (UTC)
    In that case, I could get on board with this. But before we do that, I think we have to sort out the more philosophical question of whether the policy should be written to create the appearance that it is prescriptive, or not. I didn't get a clear sense of that from your statement, but from your answer to this question, it sounds like we agree. --causa sui talk 16:51, 20 May 2009 (UTC)
    I have amended the text (new text in italics) to clarify per our discussion.Elen of the Roads (talk) 18:22, 20 May 2009 (UTC)
  • Hoaxes are not speediable. Taemyr (talk) 13:19, 24 May 2009 (UTC)
    Blatant cases of extreme hoaxes are. decltype (talk) 17:32, 24 May 2009 (UTC)
    Actually, they are speedy able under G3 as long as they are blatant - see Template:Db-hoax Excirial (Contact me,Contribs) 07:06, 25 May 2009 (UTC)
Hoax was possibly a bad inclusion, as good hoaxes can be extremely hard to spot, and for this reason a hoax should usually be the subject of a PROD or AfD. Bad, obvious, blatant hoaxes are speediable however.Elen of the Roads (talk) 09:31, 25 May 2009 (UTC)

Statement by Apoc2400 (talk · contribs)[edit]

The criteria for speedy deletion should be descriptiveprescriptive and speedy deletion should only happen within those criteria. For articles (and other) that should be deleted but do not fit the criteria, there is AfD. If AfD gets bogged down with a particular kind of articles, then we can create a new criteria. CSD is not for bypassing community consensus, and admins using it that way is a problem. Speedy deletion under IAR is only acceptable if something needs to be deleted urgently but do not fit the criteria. Never for non-urgent deletions. --Apoc2400 (talk) 18:22, 20 May 2009 (UTC)

Users who endorse this statement[edit]

  1.  M  18:31, 20 May 2009 (UTC)
  2. SoWhy 19:24, 20 May 2009 (UTC)
  3. MLauba (talk) 22:46, 20 May 2009 (UTC)
  4. ThaddeusB (talk) 23:57, 20 May 2009 (UTC)
  5.  Skomorokh  08:28, 21 May 2009 (UTC)
  6. Yep. Ale_Jrbtalk 08:51, 21 May 2009 (UTC)
  7. Davewild (talk) 21:48, 21 May 2009 (UTC)
  8. ---I'm Spartacus! NO! I'm Spartacus! 02:27, 22 May 2009 (UTC)
  9. decltype (talk) 17:35, 24 May 2009 (UTC)
  10. --Aervanath (talk) 07:23, 25 May 2009 (UTC)
  11. Juliancolton | Talk 19:22, 25 May 2009 (UTC)
  12. - Elen of the Roads (talk) 17:31, 31 May 2009 (UTC)
  13. GRBerry 14:30, 4 June 2009 (UTC)
  14. - Dank (push to talk) 14:20, 11 June 2009 (UTC)
  15. TotientDragooned (talk) 18:31, 13 June 2009 (UTC)
  16. more or less what I feel. -- Ned Scott 05:59, 29 June 2009 (UTC)


  • It sounds to me like you meant to say that the policy should be prescriptive not descriptive. Is that right? --causa sui talk 18:32, 20 May 2009 (UTC)
Yes, how stupid of me. --Apoc2400 (talk) 22:28, 20 May 2009 (UTC)
  • In other words: Leave it as is. While i respect your opinion is there any specific reasoning for this other then that we always did it like this (Along the lines: "If its not badly broken don't fix it"? Also, should we bother AFD with articles that should clearly be removed but did not fit CSD as with the cat example i described above somewhere? And last - dont get me wrong on this one - did you ever work in WP:NPP? Excirial (Contact me,Contribs) 18:59, 20 May 2009 (UTC)
  • Yes, I want to keep it as it is. I think keeping CSD strict is important for keeping admin powers limited. Bothering with AfD is not a problem unless it happens often for the same kind article, and then we can create a new speedy criteria. CSD already gets abused as it is, and we should not make it looser. Yes, I have worked at NPP, though not as much as some others. --Apoc2400 (talk) 22:33, 20 May 2009 (UTC)
  • But explain to me then why we would allow the removal of 70% or so of the created articles trough the A7 policy and don't allow the rest? Note that i am specifically refering to A7 since it is probally the most problematic one in the sense of article removals. If the system is really so vastly abused why do we allow removal of people and companies trough it while they are often notable articles? Especially since they are about the most work intensive to check for notability. Excirial (Contact me,Contribs) 23:03, 20 May 2009 (UTC)
  • I think A7 should be tightened up. Many acceptably good stubs about notable subjects do not "assert notability" in the way many CSD taggers require. --Apoc2400 (talk) 23:16, 20 May 2009 (UTC)
  • Regardless of how much you tighten up a policy, it is the execution of the policy that matters. Also, A7 is an annoying enough policy as is due to it allowing only a limited subset of articles which allows for some illogical situations. If we tighten it up even more then it would become so situational, complex and restrictive (Lets see... A7... Applies to bio's that are not adding specific accomplishments for the subject unless they refer to another wikipedia article that... etc etc etc) that i doubt it would ever be applied correctly due to the forest of exceptions. In its current form i already wonder at times if i should simply AFD anything that seems blatantly not notable just to be rid of having to track PROD templates or having to care about how literally admins follow removal rules. Most times the conclusion is that if i start thinking like this i need a break and do some WP:GEO or vandalism patrol instead :) Excirial (Contact me,Contribs) 23:47, 20 May 2009 (UTC)
  • And what happens if/when AFD gets bogged down, but there's no specific class of articles making up a significant amount of the time? Do we just say "Oh, well" and do nothing? Given that there's a dozen or so AFDs relisted every day due to lack of comments, one could argue that we're already at that point. Mr.Z-man 19:49, 20 May 2009 (UTC)
    • We have WP:PROD for that? SoWhy 19:59, 20 May 2009 (UTC)
    • Perhaps this indicates a lack of interest in actually deleting such content, rather than a need for a group of administrators to go about deleting the content. Could you provide some links to a dozen-or-less AFDs that are recent and that you think should be speedy-deleted, rather than passed on to PROD or XfD?  M  20:02, 20 May 2009 (UTC)
      • In all due right this may also be a result of patrols and admins not following the CSD policy to the letter. One editor can effectively patrol all new pages; If you factor every new page patrol into this you will end up having a filter that is as strict as each editor counted together. One thing i can assure you: If every editor followed the CSD guidelines to the letter CSD would be swamped. In the last 4 hours 319 pages were removed trough csd tags. If we can extrapolate that to a day it means that every day we remove around 1900 pages. If even 10% of these (And believe me, it IS more!) didn't follow the exact CSD guidelines we would have almost trice the number of AFD's we have now. Excirial (Contact me,Contribs) 20:34, 20 May 2009 (UTC)
        • Perhaps you could link to some examples of these?  M  20:39, 20 May 2009 (UTC)
          • Examples of what? Pages that do not follow CSD deletion guidelines to the letter? Or the source data i used for that calculation i made? Excirial (Contact me,Contribs) 20:41, 20 May 2009 (UTC)
            • Just go through a few CSD noms, and pick out some examples that are non-criteria, but should be deleted. Add a {holdon} that links here so nobody else kills them. DGG mentions that in x years and y000+ SDs, he or she has nearly never needed to stray from CSD.  M  20:50, 20 May 2009 (UTC)
              • I see multiple of those tags every day, but alas, most get removed in a short time span so o would specifically have to look for them. Unfortunately we neither have admin rights as merely going trough the A7 and G11 removals would yield enough material to support this.
As for the DGG mention, you are missing one critical actor in your analysis: Everyone, including me, technically never needs WP:IAR to remove a page. Never, ever. Why not? Simply because a page which does not meet the CSD guidelines should (par policy) be send to AFD. That is the difference: If there is a page that is certain to be removed but does not strictly fit the CSD criteria I prefer tag it as a CSD since it would be a waste of time to discuss (as I mentioned to often as an example already) an article about a pet. Technically this qualifies as usage of the WP:IAR rule. If I would choose to only tag pages that were clear vandalism I would also never need IAV. But at the same time that would mean that every other questionable article would end up at WP:AFD Excirial (Contact me,Contribs) 21:50, 20 May 2009 (UTC)
AMEN!!!!---I'm Spartacus! NO! I'm Spartacus! 00:57, 24 May 2009 (UTC)
As for what to do if AFD gets bogged down... won't happen. If an article goes through the requisite period and nobody has commented on the AFD, then the default is delete. AFD can serve as a moderated PROD. Heck, we could change the entire culture of AFD by letting people know that they don't have to !vote on every one... the only time people would need to review an AFD is if somebody says "KEEP." In which case you might be dealing with the articles creator. This shift in approach would actually reduce the burden at AFD!---I'm Spartacus! NO! I'm Spartacus! 00:57, 24 May 2009 (UTC)
In that case we should probally regroup the wikipedia deletion proces a little as there is currently quite a bit of overlap between the AFD and the PROD template. I would say CSD for the blatant cases, PROD for the articles that have a chance improving but are not really worth discussing if they don't improve (Including perhaps the A category from CSD) and AFD in case it might be worth to keep but the editor would not know for sure - i would say the non blatant notability cases could fall under this.
However, in this case we would absolutely need a means to track PROD's as their usage would vastly increase. Its easy enough to look trough my contributions every now and then and search for bluelinked CSD's since they are supposed to be removed within a day. Anything bluelinked is either a denied CSD or a removal by editor. Yet i dread having to search and look trough PROD's a week old only to see they would be contested - what then? Rollback to the old prod or just CSD it under a category called "Removed PROD" or something? Excirial (Contact me,Contribs) 07:14, 25 May 2009 (UTC)
    • To add some statistics:
Date Speedy total Relists
20 may 98 13
19 may 111 23
18 may 100 24
17 may 92 23
16 may 70 19
15 may 96 16
14 may 112 22
Note that this is a manual count - articles relisted multiple times are not counted twice because they got the template on them twice. Also the numbers might be one or two off here and there due to this being a quick tally for a general impression, rather then scientific data. What we can conclude from this? Only that about 20% of the AFD's last week are relists to establish concensus. We cannot draw a reliable comparison between AFD overload and relists yet, but it is certainly a valid hypotheses; But to lay such a link we need more information such as the amount of unique editors and the median amount of votes on an AFD every week; Is that worth the time to figure our? Possibly. Still, 20% relists is more then i expected to see. Excirial (Contact me,Contribs) 20:21, 20 May 2009 (UTC)
Just look through any recent AFD archive; probably half or more of the articles deleted were done so after unanimous decisions. Mr.Z-man 21:41, 20 May 2009 (UTC)
But the key is that they went through AFD. I would have zero problem with deleting an article where nobody commented that went through AFD. AFD/PROD serve an important role, they allow users to salvage articles or to make the case that the article should be saved. Even if the article is deleted via a snowball clause, it gives a better sense of "justice." It gives the author the chance to say, "this was more than one person making a rash decision" it gives the impression that their article had a chance.---I'm Spartacus! NO! I'm Spartacus! 02:36, 22 May 2009 (UTC)
Exactly. There are many things that are easy for one person to judge. There are many things that are not. No person is a good judge for the accuracy of his own perception. If as a community we decide what things the ordinary admin can perceive accurately, it serves as a check on the inevitable tendency to think we are cleverer than we are.
Incidentally, since it's been so often used here as an example, I've decided to stop objecting to putting real animals in A 7 along with people. It's clear from the discussion that there's strong support for adding them, and, although there are only a few a week, I think we'd spot well enough the possibly notable ones. DGG (talk) 03:52, 22 May 2009 (UTC)

Statement by SoWhy (talk · contribs)[edit]

CSD has been limited to certain cases for multiple very good reasons and the above proposal, while theoretically sound, would create a huge workload and many many conflicts. Let me outline why:

  • For the first change: Policy is a firm rule, that's why we have them. If they are really preventing urgent needed tasks, there is IAR. If they do not, there is no reason to ignore them or change this fact. I have not seen an argument that would support changing the whole idea of the way policies work and I cannot think of one.
  • CSD covers those types of articles where consensus exists that they deleted without discussion. As such, many topics have specifically been excluded because there is no such consensus (movies, books, fictional characters etc.) Saying the admins should have more leeway will undoubtedly lead to deletions of topics where we have clear consensus to not speedy-delete them.
  • Pages meeting the criteria can be removed on sight. Notice the "can"-part? No admin has to delete them, so they do not create a set of rules admins have to slavishly follow, instead they highlight our editing policy: preserve information wherever possible
  • Any generalization will increase the number of dubious calls, where admins disagree about whether something fits the criterion or not, thus leading to more conflicts. The current level of abuse of the G1 and A1 criteria by certain admins (who see them as a carté blanche) highlights this very clearly and any generalization will lead to more of this, not less, thus tying admins in debates they can easily use for more important things (at the time I'm writing this, there are 81 requests for SD and each admin who has to read or write here could instead work on those).
  • Common sense is misused as an argument too often. Wikipedia is already based on the principle that everyone is bound by the spirit of the policies, not the letter (else it'd be wikilawyering).
  • The point that the proposed change fails to see is that the strict "technicalities" are created to follow our underlying goals, i.e. that it's better to wait some days before deleting potentially valuable information rather than to speedy-delete it and lose it forever. Just because some admin thinks it's not worth keeping, it does not mean that's true. If we hadn't the huge influx of new articles, all deletions would be at WP:XFD only - and correctly so. CSD is a concession that some people flood the 'pedia with common, recurring kind of pages, mostly to promote themselves or people they know and that those cannot be handled at XFD, else it'd be backlogged too far. But that means another important thing: CSD should be the exception, not the norm. The proposal seeks to change this, albeit not that openly.
  • Ryan does fail to explain why it is a bad thing to use alternative venues like WP:PROD when an article fails the criteria. Unless there really is some reason for that, there is no reason to not rather follow the rules strictly and prod more articles than necessary if it helps to avoid the deletion of a single article that was incorrectly tagged.

That said (for the TLDR-crowd), I think there is nothing wrong with CSD at the moment that needs these kind of changes, it works quite well and 90-95% of admins in that area are able to simply prod an article that does not meet CSD criteria which in I guess 95% of all cases yields the same results without any harm. If there is a kind of article that gets created often, we can amend or create new criteria. But those changes proposed will not solve any of the few minor problems but will certainly create major new ones. New users often get bitten if their articles get deleted and if they learn that the person deleting it not even had consensus for that, they will get disillusioned. If a newbie complains to those admins, they will usually tell them why the newbie was wrong and then that's it. Saying that admins are accountable for their actions is a nice thing to say but face it, they mostly aren't. It takes huge mistakes for ArbCom to even consider de-sysopping admins who breached policies time and time again and even if they do, it will not bring back those disillusioned users who are long gone by then. No, there is no reason to change policy just because some people are uncomfortable with following it. There will always be those who do not agree with some policies but in this case the number of admins who really ignore CSD willfully is (thankfully) very small. These changes will lobby this tiny group while slapping everyone else in the face who took their time and patience to follow policy and do a good job with it. Regards SoWhy 19:59, 20 May 2009 (UTC)

Users who endorse this statement[edit]

  1.  M  20:04, 20 May 2009 (UTC)
  2. MLauba (talk) 22:48, 20 May 2009 (UTC)
  3. Hobit (talk) 02:07, 21 May 2009 (UTC)
  4. Lankiveil (speak to me) 07:06, 21 May 2009 (UTC).
  5. Yep. Ale_Jrbtalk 08:57, 21 May 2009 (UTC)
  6. Davewild (talk) 21:50, 21 May 2009 (UTC)
  7. ---I'm Spartacus! NO! I'm Spartacus! 02:33, 22 May 2009 (UTC)
  8. Too long, but good. NVO (talk) 07:48, 22 May 2009 (UTC)
  9. Attacks should be speedy deleted even if we had a hundred times as many admins and a tenth as many article writers, but otherwise I agree. ϢereSpielChequers 15:01, 22 May 2009 (UTC)
  10. decltype (talk) 17:39, 24 May 2009 (UTC)
  11. --Aervanath (talk) 07:24, 25 May 2009 (UTC)
  12. Juliancolton | Talk 19:22, 25 May 2009 (UTC)
  13. Ched :  ?  03:22, 3 June 2009 (UTC) (agree 100% with preserve what we can - but I wish there were not quite so many criteria)
  14. TotientDragooned (talk) 18:33, 13 June 2009 (UTC)
  15. S Marshall Talk/Cont 00:17, 19 June 2009 (UTC)
  16. Cybercobra (talk) 23:49, 23 June 2009 (UTC)


  • It's rather clear to me from this statement and some others that there is an epic misunderstanding about what is actually at issue here that won't go away, no matter how many times I say it. I don't know what the best way to express myself is, but it's clearly not working. --causa sui talk 20:36, 20 May 2009 (UTC)
    I think you want to say that individual administrator actions should guide policy, but one objection is that there is a consensus against certain types of administrator action. You could always write a brief comment under you statement, to clarify something.  M  20:46, 20 May 2009 (UTC)
    No, that's not what I'm saying. At all. --causa sui talk 20:48, 20 May 2009 (UTC)
  • I've read this a couple times and I can't wrap my head around your first point. You say that policies are firm rules (I suppose we need to rewrite WP:POLICY, WP:ISNOT, and WP:5) and then in the next breath you say that people can always IAR. Which one is it? --causa sui talk 20:55, 20 May 2009 (UTC)
    INo, I think he's saying that it is enough to have guidelines that set out the types of content that have a deleterious effect on the project and should be given the bum's rush. The problem is that we have guidelines masquerading as policy (these are the reasons, these are all the reasons, these are the only reasons), with a sidenote that they are not very good guidelines anyway. --Elen of the Roads (talk) 21:00, 20 May 2009 (UTC)
Policies are firm rules, no doubt about that. They represent cornerstones of our mentality, of what people think is the right way to do or what the law tells us to do. WP:IAR is policy as well but as such, it's quite clear: Do not break the rules unless you have to to improve the encyclopedia. It's like law in that regard, although I know this might draw unwanted comparisons: You are not allowed to kill people, ever. Yet there are circumstances where you are allowed to anyway, like in self-defense. Now: Is "do not kill people" not a strict rule anymore just because there might be circumstances where it is not enforced because it would be incorrect to do so? I would say no. People can IAR, but they shouldn't because if not really needed, it just creates problems (see also WP:IAR?). Regards SoWhy 21:17, 20 May 2009 (UTC)
Yeah, this is what is baffling to me about this debate. What I am reading you to be saying is that it's not okay to give people license to invoke IAR to vindicate their destructive edits. Nobody is arguing with that. IAR absolutely is not a free pass to do whatever you please, and there is nobody here who thinks it is, or that policy should be drafted in such a way that nobody thinks there will be anything done about it if they actively destroy the project or invoke IAR to do something dumb. Why this is being treated as if it were a dispute about any of those points is a complete mystery to me. --causa sui talk 21:33, 20 May 2009 (UTC)
  • With the exception of schools (and I'm not entirely sure why that is) and some certain things in userspace, I don't believe we have a "clear consensus" against speedy deleting anything. There is a big difference between "no consensus for" and "consensus against." I have, numerous time I might add, asked for some sort of evidence or logical reasoning as to how making the policy more general will lead to a rise in abuse or "dubious calls." We have had vague and general protection and blocking policies for years and do have the problems that people claim will exist with a more vague speedy deletion policy nor have there been calls for a CSD-like policy for either of those tools. From what I recall, not once has anyone provided even a crappy, evasive answer to that request; its been totally ignored every time I ask it, which only leaves me with one conclusion. I'm not sure how your argument about the spirit of policies supports what you're saying. If we only need to follow the spirit of the CSD policy, then why is people not strictly following the policy a bad thing? And then you claim that strict technicalities (which are wikilawyering and not part of the spirit of Wikipedia) are important? I don't follow. If the community doesn't want a looser speedy deletion policy, I'll accept that, but if they do, calling it a slap in the face is rather offensive. Mr.Z-man 21:34, 20 May 2009 (UTC)
    • Well, changing it against community consensus, is a slap in the face, that's what I meant. Of course, if there really is consensus for such changes, then that is how it will go. I just think from countless discussions on almost every venue that there isn't.
      As I replied above to you, the difference between blocking/protecting and deleting is visibility of those actions. The reason why WP:BLOCK and WP:PROTECT (although quite strict if actually applied the way they are written) can be less general is that both actions are visible to everyone all the time and any user can review them. A page that has been deleted for dubious reasons is gone and can only be viewed by a tiny fraction of total users (1600 admins out of what, 8 million accounts? That's less than 0.0002%), thus making it impossible for users to review those actions. Currently, dubious calls can easily be identified because it's mostly G1s and those where no criterion is named and I, amongst others, usually double-check those if I notice them because I know of the abuse.
      Not having consensus for those deletions is essentially the same as having one against them. If there is no positive consensus, it cannot be done, no matter why there isn't. That's why "no consensus" defaults to "no change" at XFD etc. You have said often that such changes are needed, yet I have not read once why alternative venues (like PROD) cannot be used. AFD has no backlog, relists are quite few and most of those could be handled by PRODs anyway. So why change CSD, a policy that is designed to be an exception to normal deletion procedure? Regards SoWhy 21:47, 20 May 2009 (UTC)
      • I'm not asking about the difference between deletion and blocking, I'm asking why admins will abuse deletion if allowed more discretion but not protection and blocking. Do you really have that little good faith toward other admins that you think just because not everyone will have the ability to review them, that admins will just go wild with deletions? Why? The amount of review that deletions can receive would not change. Are there admins who currently follow the speedy deletion policy but you think will all of a sudden start abusing it if we don't force them to work strictly by the criteria? Why would admins who've been just fine the whole time they've been on Wikipedia just suddenly start exhibiting poor judgment and then refuse to listen to complaints? Mr.Z-man 01:15, 21 May 2009 (UTC)
        • Imho some admins follow the policy just because it's so strict and once it's less strict (= when we start using words like "blatantly inappropriate" that allow interpretation), they might start deleting pages that do not fall under these criteria because interpretation and "common sense" vary very much between different admins. The MZM-drama showed us that opinions are clearly divided about some pages that one side thinks to be inappropriate for this project and the other doesn't. Now if the criteria actually used this word, any admin could delete any pages they think to be inappropriate and still follow policy, as compared to now breaking policy when doing so. It would not change the amount of review, true, but it would change the way these deletions would now be considered (by those admins) to be within policy, thus creating huge discussions about whether this is in fact the case. The current strict wording serves to limit admins also for this reasons, i.e. to avoid such discussions. I see no point in changing it to something prone to disputes. I have good faith in all admins but you and me both know that not all of them always show good judgment or listen to complaints. I do not want to make it easier for those (even though they are few) to be that way. Regards SoWhy 12:57, 21 May 2009 (UTC)
          • If you start using words like "blatantly inappropriate", which allow interpretation, then they might start deleting pages that don't fall under the criteria. When that happens, you can challenge it, and they'll have to explain why what they did was good for the project. Admins who repeatedly abuse this will face sanctions like blocks or deadminning, just like we do with everyone else who repeatedly disregards consensus and invokes IAR to justify their dumb ideas. Making the policy prescriptive doesn't save you from this; it only alters the vocabulary you use when you challenge their dumb ideas, because you challenge it on the basis that it was against policy rather than that it was bad for the encyclopedia. That's when policy stops being a useful distillation of best practices and becomes a thought terminating cliche. If it's discussion that you want to avoid, you are definitely editing the wrong encyclopedia. ArbCom realized this; Jimbo realized this; the fundamental Wikipedia policies like WP:5, WP:NOT, and WP:POLICY make it explicit. Why don't you realize it? --causa sui talk 13:11, 21 May 2009 (UTC)
            • I do not want to avoid discussions, I want to avoid drama. Every time people have gone and deleted "blatantly inappropriate" pages, there has been such drama and it wasted hours and days of precious time for nothing. Unfortunately, "bad" and "good" for the encyclopedia are vague terms and the deletionist-inclusionist battles indicate that if admins on both sides were to judge only by what is "bad for the encyclopedia", there would be a mass-deletion of allegedly inappropriate pages. It's a myth that CSD is not a useful distillation of best practices that I do not agree with and we have many strict policies that are strict for a good reason (WP:COPYVIO, WP:V or WP:CIV for example). I am all for discussing policies if the need exists to change them but I just do not see the need here and no argument yet has really shown any need for such a change. Regards SoWhy 13:38, 21 May 2009 (UTC)
              • Every time - I don't think so. Whenever there's drama over a deletion, its frequently due to a page being deleted outside of CSD, but even when it is that's not always the sole cause of the drama. Often the page itself is controversial and an AFD would have caused just as much drama. But there is not hours of drama every time a page is deleted outside of CSD. This is the FUD that I keep mentioning. That's just blatantly untrue. AFDs and normal CSD deletions are not free from drama. Again, please provides some evidence or logical reasoning as to why you think admins would just throw good judgment to the wind resulting in a "mass-deletion of allegedly inappropriate pages." I find it very difficult to believe that there are a lot of admins that know of pages that they believe are harmful to the project but refuse to do anything about it. I know I have far more respect for the project than that. WP:V and WP:CIV are not strict policies. They may be strictly enforced but neither contain the level of detail intended to be followed exactly to the letter that CSD does. You do not need to be familiar with the exact wording of any of those 3 policies to effectively follow them nor is the exact wording as critically important as it is for CSD. Mr.Z-man 14:57, 21 May 2009 (UTC)
    • We don't allow advertising, even if it doesn't influence our content. Likewise, we handle discussions in groups, even if all admins were perfect and never made mistakes when speedy-deleting. Someone going out and deleting things because they can't find enough people interested in deletion to clear the backlog... is, aside from a somewhat pov way of putting it, is not how things are done. I think that's part of the idea.  M  21:54, 20 May 2009 (UTC)
      • I'm not sure who this is replying to, but allowing the backlog to accumulate is arguably a worse option. Mr.Z-man 01:15, 21 May 2009 (UTC)
  • Regarding the PROD statement: In its current form the PROD has the incredibly annoying drawback that it may be removed by anyone for no reason whatsoever. If the creator of the article removes the prod without making a single change it means that the article will have to go trough the entire AFD procedure, therefore severely limiting PROD usability (The prod was placed for a reason, which subsequently wasn't resolved). It is something people often forget: When placing a prod it has to be monitored for removal for 7 days. When removed man needs to create an AFD anyway; One week worth of new page patrol might well end up being dozens and dozens of prods. Hence, i prefer having content pages on my watchlist rather then having to add and remove a bunch of PRODded pages. Its easy to say they are handy when not working with them frequently, but having to go trough them every day to spot template removals gets rather old after a few weeks/months. Yet an often heard comment when simply tagging for AFD (Which should be preferable par discussion on consensus here) is that it "Should just have been prodded". Excirial (Contact me,Contribs) 10:55, 21 May 2009 (UTC)
    • In my experience, 90% of all prods go uncontested and get deleted. I, too, think the removal by the creator is annoying but that's something to discuss at WT:PROD. I'd rather propose a hangon-style way where the creator can explain his reasoning on the talk page and the reviewing admin should send it to AFD as contested as long as there is any indication at all that it is not completely uncontroversial. That would sort out the removed prods (combined with an abuse filter maybe) and still not really increase workload at AFD. Regards SoWhy 12:41, 21 May 2009 (UTC)
I think if we can clarify that CSD is for items that represent a risk or a major nuisance (copyvio, libellous BLPs, gibberish), then it may open an opportunity to have a look at a better system for PROD - if the consensus is that it needs one.--Elen of the Roads (talk) 21:33, 21 May 2009 (UTC)
I have just done a check of deleted prods: 40% of them were from new articles, 60% from old ones. It serves nicely as a screen before AfD, especially because of the rule that a prodded article will be undeleted upon request. DGG (talk) 22:23, 21 May 2009 (UTC)
I find that a rather shocking percentage. The PROD category every day is on average perhaps 60 pages, which would mean about 24 pages are new articles. If you compare that to the actual amount of page removals trough CSD then we see only a tiny fraction is PROD; Seeing CSD should be pages without a change this indicates that we severely under use the prod template.
Also, i tend to agree with Elen - CSD should only be for pages that are truely unwanted, which is probally covered in the "G" section of the CSD guideline. The A section are pages that require no great speed to be removed, hence, remove A and lets just PROD them; it would also solve the annoyance of A7's limited nature. The only thing i would like is an easier way to track PROD'ded pages, and a restriction on template removal for the prod. Excirial (Contact me,Contribs) 09:27, 22 May 2009 (UTC)
A7 is the most highly used CSD criterion. Yesterday (May 22) alone, there were 310 articles deleted that included "A7" or "a7" in the deletion log out of 878 articles deleted that day. Moving all the A7s to prod would more than triple the number of prods every day and if just 5% were contested and went to AFD, it would increase the number of AFDs every day by 7-15% (assuming 100-200 AFDs every day). Mr.Z-man 17:10, 23 May 2009 (UTC)
And if just 10% of those AFD's were salvaged and provided a new user with a positive experience, then it would be well worth it. Most A7's obviously need to be deleted. A decent percentage should be declined as significance/importance is asserted. But let's use the 5% figure you reference, that would result in 100-200 new AFD's every day. Again, an AFD does not have to be commented upon to result in a delete. If nobody comments on an AFD, then treat it like a PROD and delete it. So the increased workload argument is negated. But let's say that of that 100-200, only 10% of them don't need additional work or are salvaged (as compared to salvagable.) That means that every month we are deleting 300-600 articles that we shouldn't be deleting. If just 10% of those articles were written by new users who are turned off of WP, it means that every month we are potentially chasing off 30-60 new editors with a negative first experience.---I'm Spartacus! NO! I'm Spartacus! 02:08, 24 May 2009 (UTC)

Concerning the opinion of the Arbitration Committee[edit]

Ruslik0 (talk · contribs) found the following quote from an ArbCom ruling:

The policy pages for Deletion policy, Undeletion policy, and Criteria for speedy deletion together provide policy and procedure for deletion and undeletion of pages, including pages in userspace. Administrators are expected to use their deletion and undeletion abilities consistent with these policies. Administrators who wish to delete pages that lie outside the criteria for speedy deletion should usually list those pages at the appropriate deletion discussion forum, such as Articles for deletion or Miscellany for deletion, or apply a Proposed deletion tag. This does not negate administrators' ability to delete blatantly inappropriate content even if it falls outside the formal CSD criteria, nor constrain application of our policy on biographies of living persons.

M (talk · contribs) seems very pleased with this[1], and I vastly prefer this to what is currently in the policy. That people on both sides of the debate think it supports their position is evidence that this may be the basis for a new consensus revision. I propose that the introduction should be reformed to follow ArbCom's wording. --causa sui talk 08:11, 21 May 2009 (UTC)

  • For reasons i stated above i very much prefer this explanation to the one we currently have. At the very least it clear wording that allow a tad of lenience, but at the same time denoted that CSD is nothing that should just be ignored. Excirial (Contact me,Contribs) 10:42, 21 May 2009 (UTC)
  • One factor here is the interpretation of 'usually'. Since the ArbCom is very specific as to what the exceptions are (blatantly inappropriate - so a major violation of policy or law; violations of rules on living persons) and mentions no others, I take this to be referring to those exceptions (thus avoiding a contradiction). This seems consistent with the consensus developing above against out-of-CSD deletions. I take this as support for the 'CSD are the only cases where' wording, and just two extremely specific exceptions already covered by CSD, and certainly not for any form of 'at your general discretion' interpretation.  M  17:58, 21 May 2009 (UTC)
    There is no "at your discretion" being proposed by me or anyone else and it is a complete mystery to me that you keep characterizing it that way despite my repeatedly telling you that that is not what is being proposed. --causa sui talk 10:11, 22 May 2009 (UTC)
  • I too agree that arbcom's wording was carefully chosen; we should add it to the page, attribute it to them, and that should be the only change necessary from the earlier version.So it seems we had the solution at hand from the start. if we had done some research. (sound familiar?) This does not mean, of course, that arb com makes deletion policy--they do not, and I am sure they do not think they do. However, they do properly have control over the behavior of administrators, including how closely they are expected to follow it and other policies. Their statement, is, as customary for them, somewhat ambiguous in detail, but I would leave it at that. DGG (talk) 22:15, 21 May 2009 (UTC)
    • I don't think that we should add it to the page, though I think we clearly need a references section for policy pages. ArbCom consensus on an issue should be counted like any other editor input. There is also a problem with the ArbCom wording in that it is unclear to some editors (in fact, SoWhy notes that it may be unclear). I don't think that we could have circumvented this issue, and to prevent it from arising in the future, we need to make the wording (whichever is consistent with unambiguous consensus) crystal clear and unambiguous. More below.  M  01:52, 22 May 2009 (UTC)
      • My proposal in the MZMcBride-case was to make clear how "blatantly inappropriate" is to be understood regarding to deletions, i.e. by adding a comment that this only applies when there is no controversy is even remotely possible. Unfortunately, ArbCom preferred the more unclear version in their final decision. Regards SoWhy 08:22, 22 May 2009 (UTC)

Wikipedia:Deletion policy, Wikipedia:Undeletion policy, and Wikipedia:Criteria for speedy deletion together provide policy and procedure for deletion and undeletion of pages. Wikipedia administrators are expected to use the deletion and undeletion abilities granted to them in a fashion consistent with these policies. Administrators who wish to delete articles that are clearly outside the criteria for speedy deletion should list those articles at Wikipedia:Articles for deletion or Proposed deletion. This does not negate the right of administrators to delete blatantly inappropriate content even if it falls outside the formal CSD criteria, nor constrain application of our policy on biographies of living persons.

This should provide us further context. Here too there was 16:0 unanimous consensus. Note that this is the second time in the last few months where this wording was used (tried and tested) in a case where an administrator was accused of abusing speedy-deletion. I believe this adds 16 unanimous ArbCom endorsements/votes against allowing administrators to use their judgment in cases where they find CSD to be lacking in general coverage of articles that they believe do not belong on Wikipedia.  M  01:52, 22 May 2009 (UTC)
the main difference is the change from "outside: to "clearly outside" . There were two reasons for this, I think. first, the deletions in question were very clearly outside the policy . Second, for a question of removing adminship from somebody, you want clear violations, not borderline. If they had been less totally wrong, he would have gotten a lesser penalty. We would not say that an admin should do anything he likes yup to the limit where he'd be desysopped. And RfA candidate who said avoiding desysop was the only restraint on his actions would be soundly rejected. I think we should use the first wording as the more general one. I don't think there's any actual admin here who is so reckless as to iuse that as their standard, either. There's a difference between going outside the rules a little and defying them. We're dealing here with the rules people should be keeping. Arb com will deal with what constitutes a serious violation. FWIW, I think arb com thinks policy should be prescriptive--otherwise how could they act on it? But its clear from the wording that arb com relies on us top set the limits. DGG (talk) 02:54, 22 May 2009 (UTC)
  • Actually, this is 16 arbcom members agreeing unanimously that administrators can delete blatantly inappropriate content that falls outside of the CSD criteria. Which is all I've been trying to achieve with this RFC. How about we leave the policy like it is and add this exact quote from this arbcom decision -- which you think supports your position: "This does not negate the right of administrators to delete blatantly inappropriate content even if it falls outside the formal CSD criteria, nor constrain application of our policy on biographies of living persons." I think that would definitely make this dispute go away for a long time. --causa sui talk 10:13, 22 May 2009 (UTC)
  • Could you define "blatantly inappropriate"? The way I see it, it's "blatantly inappropriate" at the level of, say, violating copyright, slander, exploiting a bug in the MediaWiki - serious, goes-beyond-content issues. What's your view? We should not add that, since it seems ambiguous, as your comments demonstrate and as SoWhy points out.  M  15:39, 22 May 2009 (UTC)
"Inappropriate" does not mean "dangerous." These words are not Wikipedia jargon, they have well-defined meanings in English. inappropriate means "not suitable for the situation" - for an article, the situation would be the encyclopedia. blatantly means "so as to be highly visible or obvious." So, "blatantly inappropriate" would mean "obviously not suitable for the situation." Mr.Z-man 15:49, 22 May 2009 (UTC)
Blatantly inappropriate means exactly that, and I disagree with your interpretation of 'encyclopedia' as the context. The context, as seems specified, is things on the order of slander against living persons (lawsuits against the foundation), and not someone's pet who you think clearly doesn't belong. I thought that the wording was clear, which is why I was optimistic before, but I was mistaken in its clarity, as is evidenced by our disagreement. You may ask ArbCom to clarify; in the meantime, it seems that this RfC remains our best option of resolving this.  M  06:20, 23 May 2009 (UTC)
The last sentence seems to be quite clearly 2 separate parts, hence the comma and the "nor" - blatantly inappropriate pages and BLP enforcement. I think you're trying to read between the lines to interpret it, when its quite clear if you just read it as normal English text. Mr.Z-man 06:25, 23 May 2009 (UTC)
Before I looked at it closely, and read it as normal English, I was convinced that it was unambiguous. Upon closer examination, and after noting your reasons, I see that it is ambiguous. I still think your interpretation is incorrect, but I think that this will be much less productive than the current RfC (should we try to get feedback on which interpretation is correct? Probably not, since this RfC addresses the point directly anyway). If you don't see it as ambiguous, I don't mind talking about that, though. Nor is just a fancy or (for negations), and a comma is normal. Flat and House; House and The Office. When a thing is in a collection of others, then the context is the type of that collection. To figure out what type of thing "House" is, you look at its pairing. So in the first, it's a domicile, in the second, it's a tv series. There is also the preceding sentence - "that are clearly outside the criteria for speedy deletion should list those". Any article that is outside the criteria, which presumably includes "obviously not suitable for an encyclopedia", should be listed. Yet, under your interpretation, the next sentence states that such content may be deleted as an administrator right. It seems clear that the type of content that should be listed is of a (non-trivially) different severity than the type that administrators have a right to delete. Maybe you don't fully agree with these two points, but do you now agree that, lacking explication, the above wording is ambiguous?  M  07:25, 23 May 2009 (UTC)
If you want a definition you might want to talk to arbcom. I don't think a definition is necessary since I think it's the appropriate level of detail; not too specific, but makes clear that it's a high standard. In the mean time I'll be speedy deleting blatantly inappropriate content with confidence since I know that arbcom is wise enough not to penalize people for doing obviously right things despite their being out of process. --causa sui talk 16:24, 22 May 2009 (UTC)
I'm curious what you think of My Works. I have declined to delete it speedily (on my talk page - it was never speedy tagged) because it doesn't fit any of the CSD categories. However, I seconded a prod, because it is not an article at all. I think if we are going to continue to have pretty spelled-out criteria for SD, then one of those criteria should be "not actually an article". LadyofShalott 02:26, 23 May 2009 (UTC)
Define "not actually an article". What if someone writes an article in verse? The introduction is certainly an article, and the poem might be moved to wikisource. But then, you could speedy it under "no context".  M  06:20, 23 May 2009 (UTC)
Given that there is a (somewhat strong) consensus developing against these sort of 'frequent out-of-csd' actions above, this doesn't seem like a good thing to do.  M  06:20, 23 May 2009 (UTC)
There is a simple solution: move it into the user space of the author (with redirect suppressed) and leave a message on the user's talk page. Ruslik (talk) 08:19, 23 May 2009 (UTC)
Agreed, there are at least 3 perfectly solid ways to get around speedy-deleting.  M  21:07, 23 May 2009 (UTC)
Welp, if someone here is vindictive enough to challenge this in front of Arbcom, we'll see who is right. --causa sui talk 09:23, 23 May 2009 (UTC)
I'm troubled by this comment - by doing this, you would not just be disobeying 'crazy rules' or some sort of vague ArbCom comment, but rather what seems to be, or is becoming, a rather strong consensus within the community that these types of deletions are inappropriate. Do you understand this, or are you saying that this does not matter, because you think ArbCom will side with you in the end once some sort of problem arises?  M  21:07, 23 May 2009 (UTC)
I don't see any point in answering this. I'll let my actions speak for themselves. --causa sui talk 21:33, 23 May 2009 (UTC)

Statement by I'm Spartacus![edit]

CSD is one of the most important areas of the project, it is also one of the least respected areas. I first got involved in it when I wrote my essay Why I Hate Speedy Deleters. 90+% of CSD'ers do a fine job and virtually all want to do what is right. Unfortunately, there are a few users who give the entire project a bad name.

I got my title for the article because despite being an admin, having close to 100 articles, an FA, and a GA to my name, I've had articles deleted---and when the are deleted it pissed me off. It usually happens because I forgot to write an article in my user space and an admin deletes the article while it is under development. This means that within a few minutes of saving a draft, my article is tagged for deletion and deleted before I had a chance to make my second save. If this pisses me off as an experienced user (and if it can happen to me as an experienced user) then we have to wonder how often this happens to less experienced users and how many of them simply leave the project never to return?

CSD is an area where a well meaning but unsupervised admins can do more damage to the project than the worst vandal.

    1. PROD is an excellent option. I rarely use PROD, but when I do the author doesn't remove the PROD tag. Why? Because when I place a PROD on an article, I notify the author that if they remove the PROD tag that their article will go to AFD. I warn them that if the article ends up at AFD that it may be deleted via a SNOWBALL or that they will spend all of their time defending it. I make my putting the PROD on it sound like a favor as it gives them a chance to salvage the article.
    2. AFD is not overworked. IMO if an article is at AFD it can be deleted if nobody comments on it when the AFD is over. The purpose of AFD isn't just to debate the merits of the article, but rather to give authors a forum to defend their article and a sense that their article received due process.
    3. If a problem existed, we would see more cases at DRV. No, that is not true. First, if the article is clearly improperly deleted, the article is likely to be recreated or the deleter to be challenged directly. The only time an article will get to DRV is if the deleter and article creator can't agree. Most of the time when a deleter is challenged the deleter will work with the creator to avoid DRV. This assumes that the creator communicates with the deleter and doesn't just walk away.
    4. The other challenge is that if a problem existed, then why don't people who argue that a problem exists take the article to DRV? Well, an article may be improperly speedily deleted, but in the current shape would fail an AFD and thus (unless improved) won't pass a DRV. Unless you are willing to salvage the article yourself, taking it DRV is a waste of time---because if you are going to improve it yourself, DRV is unnecessary! DRV is only necessary if you want to keep the article in the CURRENT shape. The reason for AFD is not to keep the article in the current shape, but rather to give people the chance to salvage it instead. One of my first articles was sent to AFD. By defending it at AFD I learned what was exepected, and the experience motivated me to get that article to pass an FA. Using DRV only matters if you want to restore the article in the current shape. The problem isn't whether or not the article should be kept/deleted, but rather does the author have a chance to defend an article? Was the article improperly deleted? Could it have been salvaged if given the chance?

Back in December I did a few "surveys" on some of the CSD criteria. IMO, 55% of the 20 A1's that I looked at should not have been deleted via CSD. In 28% of the 25 G1 cases I reviewed, CSD was incorrect. I also created a series of questions from recently deleted articles. Many of these demonstrated that CSD criteria are abused or misused. 23 of the 36 cases I highlighted had at least 20% of the respondents commenting that the article should not have been deleted via speedy deletion. If 20% of the respondents don't think the article should be speedily deleted, then I think that is a strong indicator that the article probably should not have been speedily deleted. Articles that are speedily deleted should be such that there is no doubt that even with the help of a dedicated user could not be saved. If an article can go through AFD, and be salvaged, even if in the original shape the outcome would have resulting in delete then it should not be deleted. Take this recent example of where I declined a speedy deletion on this. It was then sent to AFD where I voted to delete, but now looks like Coming Soon (2008 film). In the current shape, there is ZERO doubt that it meets our standards, in the shape it was in when I declined the speedy, I have no doubt that somebody else would have deleted it. If I found this article while reviewing the deletions of somebody else, I would have been critical of their choice to delete, but I would not have taken it to DRV. By declining the DRV, I was actually thanked for giving the author a chance to salvage it.---I'm Spartacus! NO! I'm Spartacus! 03:51, 22 May 2009 (UTC)


  • If you want to understand my motivation for writing this, this might help: recently there was a big debate on the talk page for the CSD policy over whether we need to extend the A7 criterion to include pets. Apparently someone had made an article with the content along the lines of "Sam is a horse that lives in my back yard. He likes to eat hay!", it was tagged for speedy and the speedy was declined because it didn't meet the CSD criteria.
  • Yes, it could have been sent to XFD. And there would be no reason to do that except blind, slavish adherence to rules. There would be no reason to think that the CSD had to be updated to authorize administrators to delete that kind of article except blind, slavish adherence to rules. Maybe you think admins should withhold judgment even in cases of the blindingly obvious. I won't be doing that, and neither will a great many admins, regardless of whether the policy is fixed to indicate that they can. If that creates a problem, then challenge me on my out-of-process speedy deletions. I'll be happy to explain to you why it improved the encyclopedia to speedy them. If you still object that they were out of process, you'll have to haul me before Arbcom and explain why I should be de-adminned for refusing to stop doing my job because some policy wonks at CSD dreamed up a fantasy that policy can tell me not to fix obvious problems. It's ironic because I decline about 1/4 of the speedy tags I come across, which is probably a lot more than most people. But I'm not part of this collective android dream that policy is prescriptive and I won't pretend that I don't see what I do see or that I don't understand what I do understand. It would be nice if policy would reflect common sense, but if it doesn't, it will just have to be ignored. --causa sui talk 10:24, 22 May 2009 (UTC)
    For more background, see the deletion log and deleted history of Sandwichmas. Mr.Z-man 15:12, 22 May 2009 (UTC)
Looking over that history, I would have deleted it as G3-vandalism. I wonder why Ironholds disagreed with that assessment of it. LadyofShalott 23:54, 22 May 2009 (UTC)
It doesn't really matter. I would have deleted it as "obviously doesn't belong in the encyclopedia." Sometimes -- no, most times -- you don't need policy to tell you what the right thing to do is. --causa sui talk 09:24, 23 May 2009 (UTC)
Wow, one incident where you disagree with me, and what 8 months later, that single issue is the single issue that you focus on.---I'm Spartacus! NO! I'm Spartacus! 00:02, 24 May 2009 (UTC) Oh yeah, it should be noted that the version I initially deleted, you know the one where I use the edit summary of a holiday created 9000 years ago by a jedi master? Well, that version was a vandalized version of another article, I made a mistake in deleting an article after it was vandalized without checking to see if I was deleting the original version. Once I realized that the version I deleted was not the version that was tagged (and that the vandalizer was a different person than the original creator.) I restored the article, because while the odds are that "Sandwishmas" is a hoax, hoaxes are EXPLICITLY excluded from G1 and in this case I don't see the harm in keeping the article around for a few days to find out if there is in fact a group that might have some obscure celebration called Sandwichmas. Is it probably a hoax? Yes, but there is zero need to rush it. Obviously, Z-Man thinks that the proper process is delete first, but this approach is the exact type that I find detrimental to the project and the exact reason why I think CSD'ers have such a bad reputation around Wikipedia. Too many articles are deleted by careless admins who rush to delete without considering the possibility that the articles they perceive as hoaxes/vandalism are valid. Perhaps Z-Man should familiarize himself with Note that hoaxes are generally not speedy deletion candidates. It is usually not enough for just one or two editors to investigate a hoax, as there have been cases in the past where something has been thought to have been a hoax by several editors, but has turned out to be true, and merely obscure. Suspected hoaxes should be investigated thoroughly, and only in extreme cases of blatant and obvious hoaxes should articles be WP:G3|speedy deleted as vandalism. The article should have been tagged with {{hoax}} or {{image hoax}} tag and then sent to AFD. The tag would have sent a clear message that the article was probably a hoax, but would have given the author a chance to establish its validity if in fact it were true. Believe it or not, but keeping a properly tagged article on WP for a few days while it is investigated really does the project no damage. The tag clearly would indicate that the article can't be trusted. Rushing to delete, however, can be harmful, because who knows, the article you perceive as a hoax may not be one.---I'm Spartacus! NO! I'm Spartacus! 06:48, 24 May 2009 (UTC)

Statement 2 by I'm Spartacus![edit]

Just because an article has zero chance of surviving an AFD in its current format, is not justification to Speedy Delete it per IAR.---I'm Spartacus! NO! I'm Spartacus! 06:30, 24 May 2009 (UTC)

I can agree with you on some things Spartacus, especially since we seem to have the same opinion about an extended speedy and a delayed CSD removal, but this is a bridge to far. While the current text is different WP:SNOW origionally stated "If an issue doesn't even have a snowball's chance in hell of getting an unexpected outcome from a certain process, then there is no need to run it through that process.". We use AFD to determine is an article should be kept. If there is no chance that an article is kept we should not waste time going trough a longer process even if its "Technically correct"
Policies are here to support and guide editors and should be taken seriously, but a policy itself is never a reason. A policy is written to align everyone's noses into the same direction while attempting to reach a common goal. If we are certain (Doubtless) that an action meets that goal then we should go along with it. The policy is never more important then the goal behind that policy. Excirial (Contact me,Contribs) 07:43, 25 May 2009 (UTC)
Then you are missing the reason why we have set a higher standard to CSD criteria than to AFD or PROD. CSD is for those articles who even with the help of others could not be salvaged. If an article could be salvaged, then it should not be speedily deleted. Take criteria A7, it explicitly has a lower threshold than Notability, it only asks for a claim to significance/importance. The claim doesn't have to be cited or anything, it only has to make the claim in order to avoid a CSD. Now there are some people who routinely delete articles where a claim is made with the rationale that it fails N/RS, thus doesn't meet the current guidelines. Well, CSD is intentionally set with a higher standard than AFD and those admin who delete articles with claims to importance/significance citing IAR are not following the letter or spirit of either policy. Articles that will fail AFD in their CURRENT shape, but do not fit the CSD criteria, should NOT be deleted per IAR. The reasoning is simple, just because the article will fail in the CURRENT shape does not mean that the article will fail in the shape that it might be in after a week at AFD.
I wrote, Just because an article has zero chance of surviving an AFD in its current format, is not justification to Speedy Delete it per IAR. You look at this and see only one possible outcome--if the article has zero chance of surviving an AFD in its current format then it will be deleted. You then look at SNOW and see the original words, "If an issue doesn't even have a snowball's chance in hell of getting an unexpected outcome from a certain process, then there is no need to run it through that process." Well, I look at this scenario and see TWO possible outcomes. If the article is sent to AFD, the article might fail to be improved, in which case it will be deleted. OR the article will be improved in which case it will not be deleted. In the later case, by sending the article to AFD we have educated an author on what is expected and given them a positive experience to wikipedia. If we delete the article, just because we don't see it surviving in the current shape, then we deny the author the chance to improve the article and we might lose a valuable contributor to the project. People who cite IAR/SNOW as a reason to delete articles forget that, and in your words, "The policy is never more important then the goal behind that policy." The goal behind the higher standard for CSD is not delete articles, but to afford the articles creators a chance to salvage them!
But the community has set the guidelines higher for CSD explicitly to afford authors the chance to salvage their articles. Thus, I stand firmly behind my words, Just because an article has zero chance of surviving an AFD in its current format, is not justification to Speedy Delete it per IAR.---I'm Spartacus! NO! I'm Spartacus! 08:07, 25 May 2009 (UTC)
Aaah, i see. The key words in this one was "Current form" and not "Zero chance". I agree with you on certain points now that i see the intended focus of your words. If they key is "Current form" with a might be ok notability claim then it should not be tagged for A7 in the first place; it should simply go to AFD. Yet i see two caveats: The first one is that a new editor possibly couldn't care less about HOW an article gets removed. Hence, new editors are likely as clueless about the operation CSD as they are about the operation of AFD - in turn i rarely see a newcomer even comment on an AFD and even when they do, they cannot state any policies and guidelines; Most often the comments are comparable to "But He is completely important he (And so on)". Normal responce? They just get bombarded with a bunch of guidelines and comments by other editors and the page still gets removed - And happy welcome to wikipedia! Im not sure if i would prefer a CSD template or that.
Second is how CSD currently operates. If the objective of CSD is allowing pages to improve, it seems a little weird that we delete them in minutes. Hence, my interpretation of CSD has always been that it is meant for hopeless pages that should be removed ASAP, whether PROD allows for improvement time of a week and AFD allows for community discussion and improval. Excirial (Contact me,Contribs) 10:02, 25 May 2009 (UTC)
A new user may not know the operations of how articles get deleted, but by sending it to AFD, they get to respond. At CSD, the article may be deleted within seconds of the first save. (And yes, I have see articles saved, tagged, and deleted in the same minute!) If the article is sent to AfD, they may not know how to defend the article, but the author can ask, "what do I need to do to save the article?" As I mentioned above, one of my first encounters at WP was defending an article at AFD... having people tell me that it didn't deserve an article was all the motivation I needed to get it to FA status! If that article had been speedily deleted, I might not be here today.---I'm Spartacus! NO! I'm Spartacus! 16:31, 25 May 2009 (UTC)
Good point - i remember being slightly disappointed that my first interaction with any other wikipedian was an incorrect vandalism warning - which was offset by a barnstar just a few days later. Regardless of that, i think that we should be a little more informative towards users. After the slight debacle with a page i mentioned earlier (Cozzen Publications) today it seems that we are at times overly harsh towards new users, to the extend that it is truly rude. I mean come on; 2x an A7 on a page, a level 4im for spam on a page that even I cannot see as spam, and after that banning the user for an UAA violation? I think we could soften that quite a bit by relying more on PROD then CSD for the pages that might be ok. At the very least (As you mentioned) people have a chance to see what is going on. Im actually quite intrigued by the idea of switching CSD into a "Must delete" policy for attack pages, pure spam and copyvio's (The "G's") if that idea would also involve updates to PROD. It seems to have a broad support among both camps as well - but how should that be proposed? Another comment here? Another RFC? A straw poll?
Second, would it be an idea to try to educate new patrols on how they should patrol? It seems that mainly new patrols are/were quite harsh in judging pages (Me included as well). A single vandal or unaware user just creates damage equal to a few pages and possibly leaving/getting banned at worst - a wild patrol is probably more damaging then any vandal. Hence, thats what the trouble about Huggle was about. Excellent tool, can't miss it, but deadly if not handled with care. Same with patrols. Great people they want to help, can't miss them, but o so lethal when they get trigger happy. Excirial (Contact me,Contribs) 18:25, 25 May 2009 (UTC)
Wow, a lot of questions here. First, I would remove "spam" from the list of must delete immediately. IMO those are attack pages, copy vios, vandalism, and a modified version of A7--pure vanity pages (Which is what most A7 deletions are.) Spam/advertising I think needs to be rewritten, because it is not as clear as I think it should be. I think too many articles are deleted as spam/advertising that could be salvaged, but going through CSD fails to give them a chance.
Second, before starting an RfC, IMO, you should discuss the issue on the appropriate talk pages. See if you can work out some of the details before opening the RfC or going to the Village Pump.
Third, how best to encompass change at CSD? MrZ-Man thinks that we should take people to DRV everytime we see a problem---or at least that is the approach he has advocated. I think that approach would be doomed to failure and create more drama than it is worth. I was not involved in it, but how often do we want to repeat the debacle which was the MzMcBride scenario? Taking admins to arbcom/DrV would quickly become counterproductive. In my opinion, educating new users is key. This is particularly true for new CSD'ers. While we may not be able to impact the views of the old guard who want to delete first and deal with the person behind the deletion later, I think we can have a positive effect by educating the newer users/admins. I made a mistake in trying to point out the problems of an experienced admin's deletion and realized that was like banging my head against a brick wall. My approach is to guide new CSD'ers. I try to praise people who I see doing a good job with CSD and give friendly notes those whom I think make mistakes.---I'm Spartacus! NO! I'm Spartacus! 19:04, 25 May 2009 (UTC)
Agreed, mainly because predicting "zero chance" is a very daring prediction, which is not always right. The few times (very few in recent months) I have nominated something for deletion, I have felt that there is no chance people will want to keep it, but I have been incorrect sometimes. Sjakkalle (Check!) 11:48, 31 May 2009 (UTC)

Statement by AndrewRT[edit]

I'm not an admin but I do add a fair bit of content, including over 100 new articles, and do occasionally get involved in CSD/PROD/AFD debates. I completely endorse the statement by Startacus above - "CSD is an area where a well meaning but unsupervised admins can do more damage to the project than the worst vandal." In fact I think it's worse than that - trigger happy users just nominating articles for CSD can do more damage than a vandal. I've only had one article CSD'd - one that blatently didn't meet the criteria. It infuriated me - far more than any vandal had ever done.

Given how widely CSD is misused at the moment, this proposal will make it all 100 times worse. CSD is only tolerable because it is very tightly drafted.

I'm not sure you understand what you're doing. Carry on this way, and in ten years time everyone will have given up contributing to Wikipedia and the project will be consigned to the dustbin of history. AndrewRT(Talk) 23:57, 22 May 2009 (UTC)


  • I know 150k worth of text is a lot, but you you are err... Missing the context. (Where is that A1 again :-) ). The proposals had one common basis: Simplifying or clarifying the CSD procedure. If i exaggerate CSD can currently be read as: This you can delete, that you can't and that other thing is an exception to the can/can't; That article might not be notable? Well, CSD it withing 30 minutes before it replicates and they will swarm all over the place! That article is about my cat? Prod it!Doesn't fall under A7. In other words, the policy rambles on various sides.
The suggestions on this page were meant as idea's to solve this and make CSD more consistent. One idea was to generalize the CSD (and in particular A7) criteria but force the placement of CSD template's to be stricter. EG: No category limit on A7 but a much stricter bottom at which the template can be legally placed. I assume that the combination of "CSD" near the words "Lenient" and "Common sense" set off alarm bell in quite a few people, leading to the thought that the proposal was to ease CSD restrictions in order to make article removal easier. This exact idea has been explicitly denied several times already but alas, 150k worth of words aren't that good for readability's sake.
As the discussion progressed though, it became clear that more lenience combined with stricter wording would likely be a bad idea due to the way CSD works: The more patrols, the more removals. Currently i therefore prefer to limit the CSD category to exclude A7 as they, unlike attack pages and copyvio's, don't hurt anyone if they stay. Instead those could go under a prod, with the idea that the prod would become more resilient so that it can't be just removed - one of its greatest drawbacks. The advantage of this? It allows articles a week to improve with less worry about PROD's that are just being removed which causes the need for AFD., or that a page has 30 minutes before it gets executed. At the same time it would (i hope) make the prod more "Attractive" so that some borderline or IAV + csd would be prodded instead of speedied. Excirial (Contact me,Contribs) 00:58, 23 May 2009 (UTC)
  • Yup, once again we get more mysterious remarks that this would make it easier to delete content, when it would in fact make it harder to delete content. --causa sui talk 09:25, 23 May 2009 (UTC)
    • The reason is that not everybody buys your belief that this will make it harder to delete content. Many of us look at your proposal and reach the exact opposite conclusion)---I'm Spartacus! NO! I'm Spartacus! 01:05, 24 May 2009 (UTC) (added name, had too many ~ in my sig so didn't capture user name)---I'm Spartacus! NO! I'm Spartacus! 01:53, 24 May 2009 (UTC)
      • I see no way it can be interpreted that it would be harder. The only way of making it harder will be to say you cannot speedy at all, unless the case is so clear you're prepared to invoke IAR and take your chances of being desysopped if you do one that is unsupported by the community. The present way, where you are safe if you satay within the fixed rules, take a chance of being reversed with possible unpleasant comments if you stretch them too far, and have arb com in the background if you lose sight of reality altogether, is a system under which an admin can function effectively. The main thing it needs at this point is to make it easier for non-admins to follow up and see what was actually deleted, and the ease of then taking bad decisions to deletion review. (along with a rule about not immediately deleting in-process good faith articles) DGG (talk) 01:38, 24 May 2009 (UTC)
        • Ryan, if other users reading your proposal are coming to the opposite conclusion of what you intended, then perhaps this is an indication that your proposal needs to be clearer and/or more specific as to what changes in the policy you actually intended. I re-read your proposal at the top of this page after seeing your comment above, and I still cannot understand how adoption of what you proposed would make it harder for admins to use (correctly) speedy deletion. --R'n'B (call me Russ) 20:30, 25 May 2009 (UTC)
    • Proof that this proposal is going the wrong way, and the conditions need to be tightened, not loosened

Currently at deletion review: [2], [3], and [4] , all currently unanimous for overturn. I'm going to make an analogy. If the speed limit is 65 , and 90% of people drive 75, perhaps the limit does need to be changed. If 90% drive 65, and 10% drive 75, what is needed is enforcement., not changing the law to read "you may go over the limit if you have a good reason". DGG (talk) 16:58, 24 May 2009 (UTC)

Add to that a recently A7-deleted article, now heading towards an unanimous keep at AfD. decltype (talk) 17:49, 24 May 2009 (UTC)
This is in all due rights just point pushing: If you expect flawless operation why would we even have DRV? I mean, with 800 page removals a day finding 5 DRV's is easy; Its still just a mere 0,625% of the total. I could give you several hundreds of examples where it went well. Does that mean we should be more lenient because i found more?
I already indicated i like the idea of limiting CSD and move some of those tags over to a PROD, but in all seriousness do we really need someone to breath down the neck of every new page patrol if they use CSD? And do we need to enfore guidelines in a way that would hurt the 90% driving correctly more then the 10% that drives to fast? I can guarantee you people will only slow down a certain extend before deciding they better stop driving altogether. So while you try to prevent newcomers from even [their ears nibbled] you seem to assume that it is fine to bite down hard into any editor that might even remotely and unintentionally try to nibble? Excirial (Contact me,Contribs) 08:53, 25 May 2009 (UTC)
Having said that, that does not mean i do not believe that the current guidelines need to be enforced a bit more strictly. A rather bad example i got around today is Cozzen Publications. While being created by a CoI account it looks ok, and a quick search comes up with 600 ish hits. So placing a CoI, Username and welcome tag seemed that was all that needed to be done; and perhaps a prod tag after a more extended search.
Guess what happends? One editor jumps on it, places an A7 and then out of the blue and then adds a spam4im on the creaters user page for a page that even i don't deem spam. After a removal and a message to the tagger another editor replaces the same tag. In other words... Some enforcement of the current rules would be rather nice. Excirial (Contact me,Contribs) 13:35, 25 May 2009 (UTC)
The problem is that to some editors, CSD becomes a sort of game, to play like whack-a-mole. Here's that self-promotional article, quick, a twinkle selection and off we go! Some new users start by editing something they noticed with a specific purpose, and after they're done, they look for other useful things to do. And here comes NPP, and there comes twinkle making everything as easy as the press of a button, never mind if you don't understand exactly how narrow CSD is supposed to be. Been there, done that, got the T-Shirt, and finally started to patrol from the back of the log where little emergencies exist. But I understand the mentality just too well, hence my opposition to losening the criteria even one iota. MLauba (talk) 14:50, 25 May 2009 (UTC)
Yep, a fun game till man realizes that it is always other people that must constantly play the mole. What we seem to have established is that the problem lies with the users, and in some extend with the admins. The problem is not that we do not have sufficient rules - they are just not being followed as they are supposed to be. And that is actually a major issue since all it takes to change a rule is a keyboard. Changing users is more difficult.
Additionally we are kind of in a deadlock. Blatant tags and removals would mean new users get scared away. To prevent that we try to make stricter rules which are taken with some flexibility. Enforcing them would likely scare new page patrols away which is also bad (They are after all good faith editors). So we either need to break this cycle by having stricter admins, or we should change the rules to prevent mishaps. The danger of those last two is the good old deletionist / inclusionist problem - While it seems both camps here more or less agree with the change suggested byUser:WereSpielChequers i am affraid what will happen if that one will get voted on in the open; Probally some drama at the end of the opinion spectrum from both sides. In other words: We know (At least part) of the problem, we know the cause of the problem, and yet i cannot see a sure way to solve the problem. Any other suggestions? Excirial (Contact me,Contribs) 18:49, 25 May 2009 (UTC)
Feedback & Oversight may help. For feedback, a modification of common admin tools to also post a note to a tagger's talk page when a CSD is either declined or changed because of an obvious mistagging can help educating the NPPs (plus assist in identifying admin-shopping). In the same vein, we might want to create a new template, {{oldcsd}}, similar to {{oldprodfull}} which the admin tools will also place on the article's talk page if a speedy is declined, along with the original criteria tagged and the decline reasoning (also helps preventing admin shopping). These two steps alone could help improving tagging quality as an on-going education. Regarding oversight, creating a systematic process which reviews a random sample of speedy deletion at the end of every month should provide the necessary quality control for the deleting admins.
These should work around the deadlock you describe as I don't think these adjustments in tools and processes are loaded either in an inclusionist or deletionist way. MLauba (talk) 20:00, 25 May 2009 (UTC)

Statement by Rd232[edit]

Devil's Advocate: CSD should be abolished, because it's a finger in the dyke, and it discourages fixing the dyke. We should, instead, fix the dyke: namely, address the frequency with which inexperienced contributors create not just poor first drafts of articles, but articles on topics which will never merit inclusion. The main concept here should be handholding: pushing new contributors creating articles towards using a wizard based on the articles for creation wizard, which will clarify for them what's going to be accepted and what isn't. See eg User:Rd232/Noexactmatch/Proposal. Similarly, noindexing new articles (or a subset thereof - WP:VPR#New pages, created by novices, should include noindex by default) would reduce the need for CSD, because new articles wouldn't be published via search engines so quickly. Combine the two, and the need for CSD is perhaps not eliminated, but certainly sharply reduced (limited mostly to attack/vandalism/tests). Rd232 talk 12:44, 26 May 2009 (UTC)

  • Eh, " A finger in the dyke"? I can guarantee you that every new page on wikipedia is checked by a new page patrol, and i dare bet a good deal of money that the absolute majority is checked by two or more editors. All in all we remove, say, 1600-2400 pages a day trough CSD alone and most page are removed within a couple of minutes to a few hours at worst - actually, i never even saw a CSD tag last longer then a single day for as far as i remember. In that time a search engine won't index a wikipedia page, and if i remember well all links on wikipedia already use nofollow tags so you cannot boost your links rating trough wikipedia.
    We already try to hold editors by adding welcome templates, sandboxes, personal messages and so on, but i assume a good share of the editors never even reads them. Every CSD notice template tells people to add a hangon under the existing template without removing it. Yet the majority removes the DB tag, doesn't place the hangon under the CSD tag, tries communicating on talk pages without adding hangon. etc etc etc - in short: People don't read template messages well enough. I tend to like the idea for a simple wizard that filters out CSD categories but it leaves a few problems: 1) Would the users even read and follow it? 2) Even admins and seasoned editors cannot truly agree how CSD must be executed - i think that therefore it would be extremely hard to create a wizard that can filter things out effectively. 3) Also, what prevents people from just creating a page under a different category or even directly? Would we say that everyone under 500 edits MUST use the wizard? 4) I am certain the absolute majority of the new article don't care about getting on google - instead they want to in in an encyclopedia where they are "Important". A7 is probally the most used criteria, not G11. With search engines already not having the chance to index, i doubt we would see even a minor descrese in pages created. Excirial (Contact me,Contribs) 16:29, 26 May 2009 (UTC)
OK, so there are lots of fingers in the dyke :) And a wizard is needed exactly because people ignore messages; they need to be guided through creating their first article(s). There's a fair argument for making it mandatory for newbies, but if it's not too easy to bypass the wizard and the wizard is good, we could try it first making it optional. Rd232 talk 16:44, 26 May 2009 (UTC)
Can we use another phrase than "fingers in the dyke"... I mean, this is starting to sound like Penthouse magazine (and don't prentend that I'm not the only one whose mind is in the gutter!)---I'm Spartacus! NO! I'm Spartacus! 17:33, 26 May 2009 (UTC)
I have no idea what you mean. Why would what I understand to be a pornographic magazine of some kind be interested in flood prevention systems? :)
Wikipedia is your friend: Dyke, First line (Slang) ;) Excirial (Contact me,Contribs) 19:57, 26 May 2009 (UTC)

Rd232 talk 17:40, 26 May 2009 (UTC)

He knows, ergo the smily face... probably chose the wording explicitly for the double entendre ;-)---I'm Spartacus! NO! I'm Spartacus! 20:26, 26 May 2009 (UTC)
I've had 2 or 3 articles tagged for speedy deletion. On two of those 3 cases, the article was deleted while I was making another edit. EG the article was both tagged by a user and deleted between edits! At least one of those times was at least a year after becoming an admin. The first article I had tagged was one of the first articles I wrote, but knowing that the article could be deleted in a blink of an eye sent me into a frenzy mode as I tried to deciphre wikipolicy enough to provide reason to keep the article. If it wasn't for my wife, who was an admin at the time, I might not have known how to save that article before it was deleted. Needless to say, even though that article was not speedily deleted, it left a sour taste in my mouth. As did the two articles that were speedily deleted.---I'm Spartacus! NO! I'm Spartacus! 16:41, 26 May 2009 (UTC)
Another reason to make a good wizard: avoid the pain of things being deleted, speedily or otherwise. A good wizard would dissuade people from creating junk, and help people creating good stuff. Rd232 talk 16:45, 26 May 2009 (UTC)
I still worry about how such a wizard would be rolled out. It would only be effective if we force people to use it, and the amount of topics that can be created are huge. For example, how would we phrase the check for G11: "If you want to promote yourself, yyour business, your band (etx etc etc) please don't bother writing an article." Or what about A7: "If your subject is not notable then don't write"? Any longer then 1-2 ish line questions won't get read. Any shorter and people may not even understand - also, how many people will just click trough such a izard till they see a text enter area? Excirial (Contact me,Contribs) 19:57, 26 May 2009 (UTC)
Have you seen the AfC wizard that inspired this idea? It's pretty good and shouldn't be that hard to adapt Wikipedia:Articles for creation/Wizard-Introduction. Rd232 talk 20:54, 26 May 2009 (UTC)
  • I agree, as far as deletions of 'good faith' contributions are concerned. New article textareas should be put in a column, along with a list of hints and common mistakes on the side. Actual examples would be good.  M  16:36, 26 May 2009 (UTC)
  • I disagree. Making the creation of articles more difficult discourages article creation more than the speedy deletion process does. Part of the excitement of Wikipedia is that the article pops up immediately; in many cases an approval process will leave the draft lingering in limbo for days, maybe weeks, and that won't attract anyone. If it is decided that the presence of an article needs to be "vetoed", it should be vetoed actively, not pocket vetoed by passivity. Sjakkalle (Check!) 11:52, 31 May 2009 (UTC)
  • Who said anything about approval? I think you're confused by the reference to the AfC wizard, which feeds into the AfC process. The wizard I had in mind would still lead to instant creation, but after a hand-holding process for the user. Rd232 talk 18:29, 1 June 2009 (UTC)

Statement by UninvitedCompany (talk · contribs)[edit]

There are three patterns to the way deletion process has evolved:

  1. The criteria expand over time reflecting trends at xFD and actual practice, which often pushes the boundaries purportedly set by the criteria.
  2. Over time the ratio of xFD deletions to CSD deletions declines, which means that CSD is increasingly becoming the deletion policy of the project. This is due in large measure to the fact that few Wikipedians find XFD participation to be a worthwhile form of contribution.
  3. There is a tendency over time for administrators, as a group, to delete more and more qualifying pages under any new CSD that is adopted, without assessing the value created or burden posed by each article or considering WP:BITE. Over the course of years, we reach the point where all pages matching any one of the criteria are deleted. For a recent example the copyvio and advertising CSDs, which with both interpreted quite narrowly when first adopted.

I believe these trends will continue. For this reason, a prescriptive policy is important because it is less prone to unpredictable and difficult to reverse outcomes later.

Further, CSD has evolved based on patterns at xFD. It wasn't that long ago that the only criteria for speedy deletion was "patent nonsense." To replace the CSD with some airy statement invoking the judgment of administrators would throw out this accumulated institutional memory.

Finally, I do not believe that CSD are either the most broken nor the most readily repairable of the project's deletion policies.

The Uninvited Co., Inc. 03:48, 13 June 2009 (UTC)

Users who endorse this statement[edit]

  1. Yes. Point 2 is particularly unfortunate. TotientDragooned (talk) 18:37, 13 June 2009 (UTC)
  2. Yes.—S Marshall Talk/Cont 00:14, 19 June 2009 (UTC)


  • I'm curious to know what you think are the most broken, and most readily-repairable, parts of the deletion process.—S Marshall Talk/Cont 00:19, 19 June 2009 (UTC)

Summary of outcomes[edit]

  1. CSD must be kept clear and specific, bright-line and mechanical. [1.8]
  2. Out-of-CSD speedy-deletions should instead be passed on to XfD (or PROD) for discussion, and non-compliance with this should be very rare.[1.3]
  3. The CSD do not oppose IAR, and IAR-like exemptions should not be mentioned in the CSD.[1.3,1.6]
  4. Strict CSD criteria are not bureaucracy, but instead represent the consensus against undiscussed deletions.[1.3]
  5. Though admins usually have good judgment,[1.4] higher standards apply to CSD due to less oversight,[1.8] and due to the harmful effects of deletion. Loosening the standards will make the effects worse.[1.3,1.11]
  6. Deletion, especially by one party, often has the serious harmful effects of upsetting and discouraging new contributors who are acting in good faith.[1.3]
  7. Our best-practices may be prescribed to administrators, though the descriptive/prescriptive distinction is not actually that significant an issue.[1.6,1.10,1.11]
  8. (There is some support for refactoring CSD [1.2,1.9])

(Please make any corrections to this list directly to it. Section-number 'citations' should be provided.)


Since a lot of the statements have over time evolved to separate discussions, and because the page is currently nearing 200k worth of text im writing a small conclusions / summary of what we have come up with so far. Feel free to add items but keep it short. That’s what it is called a summary for.

  1. Consensus is Key
    Consensus should be the basis of all actions and policies on Wikipedia. Natural, but discussed and agreed with.
  2. IAR and CSD
    So far it seems opinions lean to the side that CSD does not need a separate IAR clause – IAR can cover it without being explicitely included into the CSD policy.
  3. Prescriptive versus Descriptive
    There seems to be no clear consensus so far whether policies should be seen as prescriptive or descriptive, though it seems prescriptive seems to have a lead as of current. The matter to which extend this should enforced or executed is still unclear.
  4. IAR tags
    Editors seem to agree that IAR removals (Also called outside of policy removals) should be rare at best. It is still being discussed if IAR can be used on hopeless pages or if they should go to AFD.
  5. Limit CSD to blatant cases and change the PROD template so it cannot be removed by page creaters/make a hangon like system. Possibly migrate part of CSD to PROD.
    Currently seems to have a broad interest and possible consensus at least – agreements from both deletionists and inclusionists.

Excirial (Contact me,Contribs) 10:22, 25 May 2009 (UTC)

  • I don't think that some of these are accurate or supported, and some seem a bit less strict than they should be. #1 is unsupported, and may or may not be true. #2 is much stronger - this is what most of the consensus here supports: no IAR-like clauses in CSD. #3 isn't a major point. Your terminology in #4 isn't actually used on this page. The whole #5 PROD thing has not been discussed very much in the actual statements, nor is there any real consensus there. I've made a summary above that can and should (at the moment) be edited by anyone who disagrees with it, or thinks it misses an important point. Disagreements can be discussed in this comments section.  M  19:56, 25 May 2009 (UTC)
  • I couldn't find where proposal 1.9 mentioned mechanical rules in particular, so I moved it down.  M  16:49, 26 May 2009 (UTC)
Oh, OK. I'm the one that put it in - it was pretty much what I was intending to say, but it doesn't matter particularly.Elen of the Roads (talk) 22:53, 31 May 2009 (UTC)
  • It looks like you pretty much just summarized your own views and left out the opposition. --causa sui talk 22:43, 31 May 2009 (UTC)
  • Though some editors have had problems with the original (Excirial's) summary, the revised summary has received no objections, I assume it's now safe to archive and summarize this RfC.   M   21:55, 21 July 2009 (UTC)

The above discussion is closed. Please do not modify it. Subsequent comments should be made in a new section.